Alp Iiingtitutr of /Ratio

Home , Ralph Bown

VOLUME 17 MARCH, 1929 NUMBER 3

PROCEEDINGS of alp Iiingtitutr of/Ratio Engittrrni

1929 CONVENTION Washington, D. C. May 13-15

General Information and Subscription Rates onPage 404 3notitutr of AabioCligineer5 Forthcoming Meetings

FOURTH ANNUALCONVENTION Washington, D. C., May13-15, 1929

ATLANTA SECTION Atlanta, Georgia, March6, 1929

DETROIT SECTION Detroit, Mich., March15, 1929

LOS ANGELES SECTION Los Angeles, Calif., March18, 1929 NEW YORK MEETING New York, N. Y., April3, 1929

PHILADELPHIA SECTION Philadelphia, Penna., March22, 1929

PITTSBURGH SECTION Pittsburgh, Penna., March19, 1929

SAN FRANCISCOSECTION San Francisco, Calif., March20, 1929

WASHINGTON SECTION Washington, D. C., March 14,1929 PROCEEDINGS OF Xbe 3licstitute of Rabio Ettginetr5 Volume 17 March, 1929 Number 3

CONTENTS PART I Page General Information 404 Suggestions to Contributors 405 Institute Sections 406 Frontispiece, Naval Research Laboratory and Bureau of Standards 408 Institute News and Radio Notes . . 409 February Meeting of the Board of Direction...... 409 Radio Stations of Holland ...... : 410 Standard Frequency Transmissions by the Bureau of Standards410 Changes in Federal Radio Commission 411 Committee Work 411 Institute Meetings 412 Geographical List of Members Elected February 6, 1929. . 415

Applications for Membership . . 417 Officers and Board of Direction 420 Institute Committees 420

PART II Technical Papers Radio Direction -Finding by Transmission and Reception, by R. L. SMITH -ROBE 425 Some Experiments in Short Distance Short -Wave Radio Trans- mission, by J. K. CLAPP 479 Wireless Telegraphy and Magnetic Storms, by H. B. MARIS and E. 0. HIILBURT 494 Recent Developments in Superheterodyne Receivers, by G. L. BEERS and W. L. CARLSON . 501 An Extension of the Method of Measuring Inductances and Capacities, by SYLVAN HARRIS . . 506 Apparent Equality of Loudspeaker Output at Various Frequencies, by L. G. HECTOR and H. N. KOZANOWSKI 521

Facsimile Picture Transmission, by V. ZWORYKIN . 536

Notes on Grid -Circuit Detection, by J. R. NELSON . . . 551 The Radiation Resistance of Beam Antennas, by A. A. PISTOLKORS562 Discussion on Simple Inductance Formulas for Radio Coils (Harold A. Wheeler), by R. R. BATCHER and HAROLD A. WHEELER . 580

Monthly List of References to Current Radio Literature. . . 583 Contributors to This Issue ...... 589

403 3fitstitute of labia Engittrgro GENERAL INFORMATION The PROCEEDINGS of the Institute is published monthly and contains papers and discussions thereon submitted for publicatio I or for presentation before meetings of the Institute or its Sections. Payment of the annual dues by a member entitles him to one copy of each number of the PROCEEDINGS issued during the period of his membership. Subscription rates to the PROCEEDINGS for the current year are received from non-members at the rate of $1.00 per copy or $10.00 per year. To foreign countries the rates are $1.10 per copy or 011.00 per year. Back issues are available in unbound form for the years 1918, 1920, 1921, 1922, 1923, and 1926 at $9.00 per volume (six issues) or 31.50 per single issue.Single copies for the year 1928 are available at 31.00 per issue. For the years 1913, 1914, 1915, 1916, 1917, 1918, 1924, and 1925 miscellaneous copies (incomplete unbound volumes) can be purchased for $1.50 each; for 1927 at 31.00 each. The Secretary of the Institute should be addressed for a list of these. Discount of twenty-five per cent on all unbound volumes or copies is allowed to members of the Institute, libraries, booksellers, and subscription agencies. Bound volumes are available as follows: for the years 1918, 1920, 1921, 1922, 1923, 1925, and 1926 to members of the Institute, libraries, booksellers, and subscription agencies at $8.75 per volume in blue buckram binding and $10.25 in morocco leather binding; to all others the prices are $11.00 and 012.00, respectively. For the year 1928 the bound volume prices are: to members of the Institute, libraries, booksellers, and subscription agencies, 39.50 in blue buckram binding and 011.00 in morocco leather binding; to all others, $12.00 and $13.50, respectively. Foreign postage on all bound volumes is one dollar, and on single copies is ten cents. Year Books for 1926, 1927, and 1928, containing general information, the Constitution and By -Laws, catalog of membership etc., are priced at seventy-five cents per copy per year. Contributors to the PilocrEDiNGs are referred to the following page for suggestions as to approved methods of preparing manuscripts for publication in the PROCEEDINGS. Advertising rates to the PROCEEDINGS will be supplied by the Institute's Advertising Depart- ment, Room 802, 33 West 39th Street, New York, N. Y. Changes of address to affect a particular issue must be received at the Institute office not later than the 15th of the month preceding date of issue. That is, a change in mailing address to be effective with the October issue of the PROCEEDINGS must be received by not later than September 15th. Members of the Institute are requested to advise the Secretary of any change in their business connection or title irrespective of change in their mailing address, for the purpose of keeping the Year Book membership catalog up to date.

The right to reprint limited portions or abstracts of the papers, discussions, or editorial notes In the PROCEEDINGS is granted on the express condition that specific reference shall be made to the source of such mittens!. Diagrams and photographs published in the PRO- CEEDINGS may not be reproduced without making special arrangements with the Institute through the Secretary. It is understood that the statements and opinions given in the PROCEEDINGS are views of the individual members to whom they are credited, and are not binding on the membership of the Institute as a whole. All correspondence should be addressed to the Institute of Radio Engineers, 33 West 39th Street, New York, N. Y., U. S. A. Entered as second class matter at the Post Office at Menasha, Wisconsin. Acceptance for mailing at special rate of postage provided for in the Act of February 28, 1925, embodied in paragraph 4, Section 412, P. L. and It. Authorized October 26, 1927.

Copyright, 1929, by THE INSTITUTE OF RADIO ENGINEERS, INC. Publication office, 450-454 Ahnaip Street, Menasha, W18. BUSINESS, EDITORIAL, AND ADVERTISING OFFICES, 33 West 39th St., New York, N. Y.

404 SUGGESTIONS FOR CONTRIBUTORS TO THE PROCEEDINGS

Preparation of Paper Form-Manuscripts may be submitted by member and non-member contributors from any country.To be acceptable for publication manuscripts should be in English. in final form for publication, and accompanied by a summary of from 100 to 300 wordsPapers should be typed double space with consecutive numbering of pages. Footnote references should be consecutively numbered, and should appear at the foot of their respective pages. Each reference should on ntain author's name, title of article, name of journal, volume page, month, and year.Generally, the sequence of presentation should be as follows: statement of problem; review of the subject in which the scope, object, and conclusions of previous investigations in the same field are covered; main body describing the apparatus, experiments, theoretical work, and results used in reaching the conclusions conclusions and their relation to present theory and practice; bibliography. The above pertains to the usual type of paper. To whatever type a contribution may belong, a close conformity to the spirit of these suggestions is recommended. Illustrations-Use only jet black ink on white paper or tracing cloth.Cross-section paper used for graphs should not have more than four lines per inch.If finer ruled paper is used, the major division lines should be drawn in with black ink, omitting the finer de visions.In the latter case, only blue -lined paper can be accepted.Photographs must be very distinct, and must be printed on glossy white paper. Blueprinted illustrations of any kind cannot be used.All lettering should be 1/f. in. high for an 8 x 10 in. figure. Legends for figures should be tabulated on a separate sheet, not lettered on the illustrations. Mathematics-Fractions should be indicated by a slanting line. Use standard symbols. Decimals not preceded by whole numbers should be preceded by zero, as 0.016. Equations may be written in ink with subscript numbers, radicals, etc., in the desired proportions. Abbreviations-Write a.c. and d.c., ko, pf, opt, emf, mh, ph, henries, abscissas, antennae Refer to figures as Fig. 1, Figs. 3 and 4, and to equations as (A). Number equations on the right, in parentheses. Summary-The summary should contain a statement of major conclusions reached, since summaries in many cases constitute the only source of information used in compiling scientific reference indexes.Abstracts printed in other journals, especially foreign, in most cases consist of summaries from published papers.The summary should explain as adequately as possible the major conclusions to a non -specialist in the subject. The summary should contain from 100 to 300 words, depending on the length of the paper.

Publication of Paper Disposition -All manuscripts should be addressed to the Institute of Radio Engineers. 33 West 39th Street. New York City. They will be examined by the Committee on Meetings and Papers and by the Editor.Authors are advised as promptly as possible of the action taken, usually within one month. Proofs-Galley proof is sent to the author. Only necessary corrections in typography should be made. No new material is to be added. Corrected proofs should be returned promptly to the Institute of Radio Engineers, 33 West 39th Street, New York City. Reprints-With the notification of acceptance of paper for publication reprint order form is sent to the author. Orders for reprints must be forwarded promptly as type is not held after publication.

405 INSTITUTE SECTIONS Chairmen Secretaries ATLANTA Walter Van Nostrand George Llewellyn, P. 0. Box 1593, Atlanta, Ga. BOSTON George W. Pierce Melville Eastham, 30 State St., Cambridge, Mass. BUFFALO -NIAGARA L. C. F. Hoyle P. S. March, 428 Richmond Ave., Buffalo, N. Y. CHICAGO H. E. Kranz J. H. Miller, Jewell Electrical Inst. Co., 1650 Walnut St., Chicago, Ill. CLEVELAND Bruce W. David D. Schregardus, Ohio Bell Tel. Co., 750 Huron Road, Cleveland, Ohio CONNECTICUT VALLEY Q. A. Brackett F. C. Beekley, 96 S. Main St., W. Hartford, Conn. DETROIT Earle D. Glatzel W. R. Hoffman, 615 West Lafayette Blvd., Detroit, Mich. LOS ANGELES Thomas F. McDonough W. W. Lindsay, Jr., 1348 Club View Drive, Los Angeles, Calif. NEW ORLEANS Pendleton E. Lehde Anton A. Schiele, 1812 Masonic Temple, New Orleans, La. PHILADELPHIA J. C. Van Horn John C. Mevius, 5135 N. Fairhill St., Philadelphia, Pa. PITTSBURGH W. K. Thomas A. J. Buzzard, 915 Penn Ave., Pittsburgh, Pa. ROCHESTER A. B. Chamberlain A. L. Schoen, 3 Kodak Park Rochester, N. Y. SAN FRANCISCO Donald K. Lippincott Paul Fenner Custom House, San Francisco, Cal. SEATTLE W. A. Kleist Abner R. Willson, 8055 -14th Ave., N. E., Seattle, Wash. TORONTO A. M. Patience C. C. Meredith, 110 Church St., Toronto, Ontario WASHINGTON F. P. Guthrie Thomas McL. Davis, (Acting Secretary), 4302 Brandywine St., N. W., Washington, D. C.

406 i

, GENERAL VIEW OF NAVAL RESEARCH LABORATORY, ANACOSTIA, D. C. (Several additional buildings have been added since this photo was taken.)

AIRPLANE VIEW OF BUREAU OF STANDARDS, WASHINGTON, D. C. Radio Building at right in background.

INSPECTION TRIPS AT CONVENTION Two full afternoons of the Fourth Annual Convention will be devoted to inspection trips to the Naval Research Laboratory and the Bureau of Standards, in Washington. The Naval Research Laboratory is located on the Potomac River about eight miles from the center of Washington. It will be reached by buses going via the White House, Treasury. Building, up historic Penn- sylvania Avenue, past the Capitol, Botanical Gardens, Congressional Library, Navy Yard, and St. Elizabeth's Hospital. The trip to the Bureau of Standards will be made by buses also. The route is to include the fine residential section of Washington, and a four - mile ride through Rock Creek Park. Returning, the convention delegates will cross the Million Dollar Bridge, past the former home of Herbert Hoover, the "S" Street home of Woodrow Wilson, and many other historic places of interest. . INSTITUTE NEWS AND RADIO NOTES

February Meeting of the Board of Direction The February meeting of the Board of Direction of the Insti- tute, held on the sixth of the month in the offices of the Institute in New York City, was attended by the following Board members: A. Hoyt Taylor, President;Melville Eastham, Treasurer; Arthur Batcheller, R. A. Heising, J. V. L. Hogan, L. M. Hull, C. M. Jansky, Jr., R. H. Manson, L. E. Whittemore, and J. M. Clayton, Secretary. The following were transferred or elected to higher grades in the Institute: Transferred to the Member grade, Harold Gray; elected to the Member grade, Lieut. H. N. Coulter and Mrs. J. D. Stewart. One hundred and six Associate members and ten Junior members were elected. With great regret the Board of Direction acceded to Dr. Goldsmith's request that he should not be reelected Editor of Institute Publications, a position which he has held for the past sixteen years. By unanimous vote of those members present the following resolution was adopted by the Board: WHEREAS:The members of the Board of Direction of the In- stitute of Radio Engineers are deeply appreciative of the in- valuable services rendered by Mira Dorton eolbomitb as Editor of thePROCEEDINGSsince the formation of the Institute; and WHEREAS:This contribution made by him as a pioneer in the development of the radio art is immeasurable in the light of future progress; and WHEREAS:The high standards and ethics adopted by him as Editor have earned for thePROCEEDINGSan enviable position in the field of radio engineering literature; Therefore Be it Resolved: That the Board of Direction, as individuals and as a whole, express to Dr. Goldsmith deep regret at the termination of his tenure of office as Editor, and profound appreciation of his untiring efforts and conscientious zeal, to which thePROCEEDINGSfor the past sixteen years shall always remain a fitting testimony. To succeed Dr. Goldsmith as Editor of Publications and Chairman of the Board of Editors, the Board appointed Dr. Walter G. Cady of Scott Laboratory, Wesleyan University, 409 410 Institute News and Eadio Notes

Middletown, Connecticut.Dr. Cady served on the Board of Direction for 1928 and was the recipient of the 1928 Morris Liebmann Memorial Prize. He is well known as a professor of physical science, a research worker in piezo electricity, a past associate editor of the Physical Review, and a contributor on numerous occasions to the PROCEEDINGS.

Radio Stations of Holland From the Dutch Government a revised list of high -frequency assignments in Holland (1,500 kc and above) has been received by the Institute office.This is a revision of the information re- lating to Dutch stations contained in the list of high -frequency stations published in the November, 1928, issue of the PRO- CEEDINGS. Copies of the revision are available to members upon request at the Institute office.

Standard Frequency Transmissions by the Bureau of Standards The Bureau of Standards announces a new schedule of radio signals of standard frequencies, for use by the public in calibrating frequency standards and transmitting and receiving apparatus. This schedule includes many of the border frequencies between services as set forth in the allocation of the International Radio Convention of Washington which went into effect January 1, 1929. The signals are transmitted from the Bureau's station WWV, Washington, D. C. They can be heard and utilized by stations equipped for continuous -wave reception at distances up to 1,000 miles from the transmitting station. The transmissions are by continuous -wave radiotelegraphy. The modulation which was previously on these signals has been eliminated. A complete frequency transmission includes a "general call" and "standard frequency" signal, and "announce- ments." The "general call" is given at the beginning of the 8 -minute period and continues for about 2 minutes. This includes a statement of the frequency: The "standard frequency signal" is a series of very long dashes with the call letter (WWV) inter- vening. This signal continues for about 4 minutes. The "an- nouncements" are on the same frequency as the "standard frequency signal" just transmitted and contain a statement of the frequency.An announcement of the next frequency to be transmitted is then given.There is then a 4 -minute in- Institute News and Badio Notes 411 terval while the transmitting set is adjusted for the next frequency. Information on how to receive and utilize the signals is given in the Bureau of Standards Letter Circular No. 171, which may be obtained by applying to the Bureau of Standards, Washington, D. C. Even though only a few frequency points are received, persons can obtain as complete a frequencymeter calibration as desired by the method of generator harmonics, information on which is given in the letter circular. The schedule of standard frequency signals is as follows:

RADIO TRANSMISSIONS OF STANDARD FREQUENCY; SCHEDULE OF FREQUENCIES IN KILOCYCLES Eastern Standard Time March 20 April 22 May 20 June 20 July 22 10:00-10:08 P. M. 1500 4000 125 550 1500 10:12-10:20 1700 4500 150 600 1700 10:24-10:32 2250 5000 200 700 2000 10:36-10:44 2750 5500 250 800 2300 10:48-10:56 2850 6000 300 1000 2700 11:00-11:08 3200 6500 375 1200 3100 11:12-11:20 3500 7000 450 1400 3500 11:24-11:32 4000 7300 550 1500 4000

Changes in Federal Radio Commission The Institute is honored in the action of the President of the United States in appointing to the Federal Radio Commission two of its Board members: Arthur Batcheller and C. M. Jansky, Jr. At this writing these appointments have yet to be confirmed by the Senate. Mr. Batcheller is to succeed 0. H. Caldwell from the first zone, and Professor Jansky is to succeed Sam Pickard from the fourth zone. Mr. Batcheller was appointed to membership on the Board of Direction in 1928, and was elected to Board membership by the Institute for a three-year term beginning with 1929. Professor Jansky was elected to membership on the Board of Direction by the Institute in the last election. The Senate Committee has confirmed the reappointment of Judge Ira E. Robinson of the second zone, Judge Sykes of the third zone, and Mr. H. A. Lafount of the fifth zone.

Committee Work 1929 COMMITTEE CHAIRMEN At the February 6th meeting of the Board of Direction the Board approved the appointment of committee personnel for 1929, the chairmen of the various committees being as follows: 412 Institute News and Radio Notes

Board of Editors, W. G. Cady; Committee on Nominations, Melville Eastham; Committee on Broadcasting, L. M. Hull; Committee on Sections, E. R. Shute; Committee on Admissions, R. A. Heissing; Committee on Constitution and Laws, R. H. Marriott; Committee on Membership, I. S. Coggeshall; Com- mittee on Standardization, J. H. Dellinger; Committee on Meet- ings and Papers, K. S. Van Dyke; Committee on Publicity, W. G. H. Finch; Committee on Institute Awards, Melville Eastham. COMMITTEE ON ADMISSIONS A meeting of the Committee on Admissions was held on February 6th in the office of the Institute in New York City. The following members were present: R. A. Heising, chairman; H. F. Dart, George Lewis, C. M. Jansky, Jr., and E. R. Shute. The committee considered six applications for transferor election to the higher grades of membership in the Institute. Institute Meetings PROPOSED LEHIGH VALLEY SECTION Much interest has been expressed on the part of the Institute members residing in the vicinity of Allentown, Pa., in the organization of a Lehigh Valley section of the Institute. B. H. Eckert and F. J. Hardner have been instrumental in thepre- liminary organization plans which brought fortha meeting of the Institute members at Muhlenberg College, Allentown, Pa. on February 8th, 1929. At this organization meeting B. H. Eckert and C. F. Maylott presided. The paper by R. L. Smith -Rose, "Radio Direction -Finding by Transmission and Reception," was presented by Mr. Maylott. F. J. Hardner presented a paper, "Radio Noise Finding and Its Elimination."Messrs. Maylott, Hardner, Thomas, Muthard, and Kleck participated in the discussion which followed the presentation of these papers. Mr. Maylott was elected chairman of the temporary organization. Another meeting is to be held on the 8th of March. NEW YORK MEETING At the New York meeting of the Institute held on February 6th in the Engineering Societies Building, 33 West 39th Street, A. Hoyt Taylor was inaugurated as President of the Institute. Ralph Bown opened the meeting with a short speech of introduction to Dr. Taylor. Institute News and Radio Notes 413

A paper by R. L. Smith -Rose, of the National Physical Laboratory, Teddington, England, on "Radio Direction -Finding by Transmission and Reception" was presented by L. M. Hull, of the Radio Frequency Laboratories, Boonton, New Jersey. Following the presentation of the paper the following participated in its discussion: A. Hoyt Taylor, L. M. Hull, Lester Jones, S. W. Dean, H. E. Hallborg, Stuart Ballantine, and others. Three hundred and twenty-five members of the Institute and guests attended this meeting. The April New York meeting, which is to be held on April 3rd, will consist of a symposium on frequency measurements. There will be seven contributions to this symposium by engineers in- timately associated with frequency measurements and standardization. BUFFALO -NIAGARA SECTION L. Grant Hector, of the University of Buffalo, presented a paper, "Apparent Equality of LoudspeakerOutput at Various Frequencies", at the January 17th meeting of the Buffalo - Niagara Section. L. C. F. Horle, chairman, presided. Messrs. Horle, Lidbury, Henderson, Stone, and others participated in a discussion which followed. Twenty-five members of the section attended the meeting.This paper is printed elsewhere in this issue of the PROCEEDINGS. CLEVELAND SECTION On January 25th a meeting of the Cleveland section was held jointly with the Cleveland Astronomical Society in the Case School of Applied Science, Cleveland, Ohio, presided over by Chairman Bruce W. David and attended by one hundred and sixty members and guests. A paper was presented by H. T. Stetson entitled "Sunspots, Radio, and Weather." The paper showed the striking relations between solar activity and certain of the earth's meteorological conditions. Radio reception, the earth's magnetic field, and to a certain extent the weather are affected by sunspots.The paper was illustrated with lantern slides. Los ANGELES SECTION The Los Angeles section held a meeting in the Elite Cafe, Los Angeles, on January 21st, attended by one hundred and twenty-five members and guests. T. F. McDonough, chairman of the section, presided. 414 Institute News and Radio Notes

A paper entitled "The Nature of Sound, Its Transmission and Reproduction" was presented by A. P. Hill, of the Southern California Telephone Company.

ROCHESTER SECTION The Rochester section held a meeting in the Sagamore Hotel, Rochester, N. Y., January 11th, presided over by E. C. Karker, vice-chairman of the section. Joseph P. Maxfield, of the Victor Talking Machine Co., presented a paper on "Physical Requirements of High Quality Audio -Frequency Reproduction."By means of lantern slides and a special orthophonic victrola, the speaker demonstrated improvements that had been made recently in the art of sound reproduction.The limitations of mechanical and electrical reproduction of sound were pointed out. This was a joint meeting with the Rochester section of the American Institute of Electrical Engineers and the Rochester Engineering Society.

SAN FRANCISCO SECTION A meeting of the San Francisco section was held in the Hotel Bellevue, San Francisco, January 23rd, attended by twenty-eight members. Leonard F. Fuller, chairman, and Donald K. Lippincott, vice-chairman, presided. Harry R. Lubcke presented a paper on "Vacuum -Tube Voltmeter Design." The election of officers was held with the following results: Donald K. Lippincott, chairman; Walter D. Kellogg, vice-chairman; Paul R. Fenner, secretary -treasurer. Proceedings of the Institute of Radio Engineers Volume 17, Number 3 March, 1929

GEOGRAPHICAL LOCATION OF MEMBERS ELECTED FEBRUARY 6, 1929

Transferred to the Member grade Ohio Cleveland, WJAY, Schofield Bldg. Gray, Harold E.

Elected to the Member grade Dist. of ColumbiaWashington, 408 International Bldg. Stewart, Jessica Dee Nicaragua Managua U.S.S. Rochester, Balboa, C. Z Coulter, Howard N

Elected to the Associate grade Alabama Mobile, c.o Reynolds Music House Co... Helt, Sanford Mobile, P. 0. Box 801 Helt, Scott California Oakland, 1208 E. 18th St Brearty, Lawrence S. Los Angeles, 444 Kings Road Groelean, G. M. San Francisco, 2370 Filbert St Smith, Herbert H. San Jose, 398 N. 8th St Kelliher, John Edward Colorado Colorado Springs, 920 E. Monument Harvey, Francis M. Dist. of ColumbiaWashington, Bureau of Standards Brattain, Walter H. Washington, 3706 -24th St., N. W Day, John F. Washington, 401 -23rd St., N. W Detwiler, Donald J.

Illinois Chicago, 4134 North Richmond Ave.. . . . Bross, Fred 0. Chicago, 2787 Frances Place Edwards, Charles Chicago, 6257 Harper Ave Hermanuts, Ray J. Chicago, 2251 E. 71st St Timmings, George H. Oak Park, P. 0. Box 552 Hoffman, Walter Henry Iowa Blairsburg Baughman, Raymond C. Kansas Atchison, 624 U Sc Gier, Willard Marion Massachusetts Arlington, 68 Marathon St Dempsey, John P. Boston, P. 0. Box 1678 Decker, Donald Philips Cambridge, 15 Holworthy Hall Hunkins, Harold R. Cambridge, 112 Lakeview Ave Noyes, Atherton, Jr. Cambridge, Harvard University, Holworthy 16 Packard, Alden C. Cambridge, 15 Holworthy Hall Taylor, John Pratt Medford, 21 Frederick Ave Surette, Dennis C. N. Wilbraham, Cottage Ave Garvey, Edmond Springfield, 407 St. James Ave Chapman, Alan B. Springfield, 19 Elliott St Van Doeren. C. A. Michigan Detroit, 11338 Dexter Blvd Huber, John E. L'E. Detroit, 504 Commerce Bldg Kratokvil, Frank M. Detroit, 15364 Oakfield Ave Martin, Robert D. Detroit, 18508 Welland Ave .O'Neil, John J. Detroit, 135 Nevada. W Porter Samuel Jackson, 1019 First St Rich, a. E. Jackson, 434 Stewart Ave Riethmiller, Earl R. Missouri St. Louis, 6339 Marquette Ave. Riddle, Ruston L. Nebraska York, 1214 Lincoln Ave Meyer. Albert New Jersey Allendale, P. 0. Box 207 Asten, Oliver B. Atlantic City, 225 Atlantic Ave , Apt. B-6 Enderle, Jackson J. Irvington, 617 Grove St. Hart, James J. Linden, 1104 Wood Ave Holets, Alexander C. New York Brooklyn, U Wyckoff St Kirdahy, Emil Buffalo, 15 University Ave Schwing, Ruelell L. Great Neck 20 Arrandale Ave Genner, W. Gordon, Jr. New York City, 336 East 5th St Barnueba, Richard New York City, 116 Broad St. Dana, Hermann New York City, 273 West 113th St. Dias, Ernest New York City, 120 East 30th St Hastings, Gerald M. New York City, Elec. Research Products, Inc , 250 W. 57th. Hipp, W. S., Jr. New York City, 463 West St., Room 279 Kurtinaitis, John V. New York City, 152 D okman St Murphy, E. Edward New York City, Elec. Research Products, Inc , 250 W. 57th Padden, Cecil John New York City, 269 W. 34th St. Payette, Walter 13. Riverhead, L.I., c/o Radio Corporation of America Henery R. S. Schenectady, 221 Seward Place Biver, Carl J. Schenectady, 2 Eagle St. Frink, Frederick W. 415 416 Geographical Location of Members Elected February 6, 1929

Schenectady, 607 Chapel St. Lynn, Roland A. Schenectady, Y.M.C.A. Bldg McLennan, Miles Ayrault Schenectady, 614 Campbell Ave. Richards Philip A. Schenectady, Y.M.C.A., Room 425 Tanke, Harold F. North Carolina Norlina, Norlina Hotel Sehulke, Enno Ohio Akron, 179 Ido Ave Smith, Paul C. Cleveland, 4124 Bailey Ave Deutsch, E. J. Cleveland, 3905 Svoboda Ave Pitonak, Joseph C. Cleveland, 2845 Prospect Ave UaThreat, Stanton Lakewood, 1288 Ramona Ave Drummond, P. Marion, 360 Silver St Ackerman, Francis R. Newphiladelphia, 334 Minnie!) Ave Nolan, John Salem, 804 McKinley Ave Ludlow, Gilbert H. Warren, 1903 Grove Place Wason, R. E. Oklahoma Tulsa, Radio Dept., Skelly Oil Co Rand, George L. Tulsa, c/o Dodge Electric Co., 318 So. Boulder .Swan, Merrill LeRoy Pennsylvania Allentown, 959 Turner St Keck, Kenneth K. Boyertown, 36 North Reading Ave Markle, J. E. Philadelphia, 4916 Chestnut St Kienzle, D. R. Wayne, 225 W. Wayne Ave. Adelberger, Paul J. South Dakota Huron, Box 683 Griffiths, A. Elmer West Virginia Kingwood, 134 Prior) St. Michelson, Peter Tennessee Saltillow Berry, James S. Texas Houston, 403 Studewood Harold, George Virginia Quantioo, Poet Radio Station, Brown Field ..Beardsley, Franklin Washington Seattle, 4055 -42nd Ave., S. W Mood, George T. Wisconsin Fond du lac, 284 Dixie St Eckert, Walter R. Madison, 1530 Jenifer St Hagen, Paul A. Milwaukee, 672 Grove St Frohrieb, Edward F. Canada Hamilton, Ont., 339 Wilson St Clemnts, B. D. Hamilton, Ont., 267 Cannon St., E Crofte, Cecil T. Hamilton, Ont., 85 Garfield Ave 8 Bain, J. R. Penticton, B. C Howse, J. Toronto, Ont., 67 McLean Blvd. Clarkson, Percy W. Toronto, Ont., 578 Spadina Ave Freeman, William Toronto, Ont., 41 McPherson Ave Glassford, J. 0. Toronto 12, Ont., 1331 Avenue Road Kinnear, Donald R. Toronto, Ont., 578 Spadina Ave Southern, Sheraton Toronto,Ont.,OntarioForestryBranch, Parliament Bldg Ward, Charles Channel Islands Guernsey, 16 Bordage St Laker, Edward W. China Canton, Sun Yatsen University Chu, Chill Teih Denmark Copenhagen, Vennemindevej 3, 3 Str Svenningeen, Karl England Branton, Lincoln Sparrow, A. W. Cambridge, Corpus Christi College Landale, S. E. A. London, Berkeley St., e/o Thomas Cook, Ltd-Samarasokara, V. R. London, Wl, Berkeley St., o/o Thomas Cook, Ltd. Wijeyeratne, P. de S. Swindon, Wilts, 32 Prospect Hill Mustohin, Norman Whitley Bay, Northumberland, 25 Marden Road South Gow, Charles N. Germany Berlin, Wilmersdorf, Hildegardstr. 13 b Lock, M. J. Heilbronn, Bookingen Hertweck, C. New Zealand Dunedin, 418 Anderson's Bay Road MoKewen, J. D.

Elected to the Junior grade California Los Angeles, Y.M.C.A., 715 Hope St Belleville, Logan Colorado Denver, 1237 Elisabeth St. Foraker, William Nelson Illinois Chicago, 5211 Kimbark Ave MoCammon, Donald Paw Paw Braffet, Donald H. Kansas Coffeyville, P. O. Box 190 Russell, Charles New York Buffalo, 17 William St Freedman, Sam Schenectady, Y.M.C.A., State St. Lee, Samuel T., Jr. Ohio Lakewood, 12974 Emerson Ave Thompson, J. Kent Pennsylvania Philadelphia, Mt. Airy, Thouron Ave. and Gorges Lane Ellinwood, Kenneth M. England London N 5, 59 Highbury, New Park O'Rourke, Sydney P. Proceedings of the Institute of Radio Engineers Volume 17, Number 3 March, 1929

APPLICATIONS FOR MEMBERSHIP Applications for election to the various grades of membership have been received from the persons listed below, and have been approved by the Committee on Admissions. Members objecting to election of any of these applicants should communicate with the Secretary on or before March 29, 1929.These applicants will be considered by the Board of Direction at its April 3rd meeting.

For Election to the Associate grade California Los Angeles, c/o California Radio Service, 800 N. Spring St. Herrnfeld, Frank P. Los Angeles, 5308 Third Ave. Howe, Franklin Joachino Los Angeles, 521 Amethyst St Oodrys, Arthur Palo Alto, Federal Telegraph Co. Shermund, Ralph C. San Joee, 1148 Lincoln Ave Seine. Harry Stanford University, Box 1331 Skinner, Clifton Ross Connecticut Bridgeport. 68 Mead St Tillman, John Elmer Kensington, P.O. Box 276 Edgerly, A. H., Jr. Dist. of ColumbiaWashington, Radio Section, Bureau of Standards. Arnold, Prescott N. Washington, 1705 Bay St.. S E Bower, Ward E. Bellevue, Naval Research Laboratory Hyland, Lawrence A. Washington. 2301 Cathedral Ave., N.W. Jackson, William E. Washington, 900 F St., N.W. Kaplan, Jacob Washington, 2618 -13th St. N.W.. Moulton, Theodore S. Washington, TheChastleton, 1701 16th St. N.W.Newton, Jane Elisabeth Illinois Chicago, 5054 Agatite Ave Beusman, Robert Marshall Chicago, 6716 Parnell Ave. Cubert, Joseph R. Chicago, 411 S. Ashland Blvd. Kempf, Frank J. Chicago, 3250 North Crawford Ave Kincaid, Owen D. Chicago, 501 North Central Ave Knutson, Ben M. Chicago, 1706 N. Mayfield Ave. Lamb. Harold A. Chicago, 3823 S. Kedsie Ave Patterson, Joseph A. Chicago, 3621 E. 107th St. Siovic. John E. Chicago, 7728 Calumet Ave Van Sickel. R. E. Lake Forest. 1199 Edgewood Road...... Baker, James Maurice Litchfield, Box 214 Hopper, C. L. Indiana Evansville, 112 Jackson Ave Norwood, VandleClarence Iowa Des Moines, 1538 -31st St Hutchison, Sam T. Kansas Pittsburg, Pittsburg Radio Service Forman, H. W., Jr. Louisiana New Orleans, 2625 Jefferson Ave Goldstein, Henry R. New Orleans, 4614 Carondelet St...... Adler, Leonard E. Maine Yarmouth, Greeley Road Tamminen, Nestor Massachusetts Belmont, 98 Payson Road Anthony, A. W., Jr. Boston, 472 Massachusetts Ave Nunes, F. J. Boston. Custom House. c /o Supervisor of Radio Weston, Irving L. Cambridge, 22 Bigelow St Battison, Wallace 4. Cambridge, General Radio Co., 30 State St Thiessen, Arthur E. Chatham, 0/0 Radio Corporation of America Thing. K. North Attleboro, 9 Elm St. Hale, Willis L. Revere. 62 Malden St. Fletcher, Frederick R. Salem, 281 Grove St. Heffernan, S. K. Springfield, 62 Kimberly Ave Ferguson, George Well Stockbridge, Box 982 Harrington, George N. Michigan Alpena. 111 Tawas St. Snyder, Leslie T. Ann Arbor. 1030 Church St Mathison, Gerald Detroit, 3028 Lothrop Ave Woehr, William

Minnesota Minneapolis, 1521 University Ave., S. E . Reed, Harry F. Missouri Bucklin Holmlund. A. Earle Davis. Glen A. Mendota, R.F.D St. Louis, 4047 W Pine St Humphreys, Irl W. Nebraska Falls City, 1801 Morton St Cheeky, Arthur D. New Hampshire Claremont, 227 Main St Hodge, V. W. New Jersey Bloomfield, 25 Grace St Henry, T. J. Camden, 303 North 6th St Galvin, Robert Clayton, 326 Broad St . Purvis, Charles G. Deal, Box 122 Shaw, Robert C. Glen Ridge, 129 Midland Ave Grabo, Irving C. 417 418 Applications for Membership

Linden, 104 Luttgen Place Angle, Gafford Brock New Brunswick, R.C.A. Radio Station Bohman. Victor A. New Brunswick, 149 Hale St Lucas, Earle F. Stanhope, Box 261 Peterson, Arthur C. New York Astoria, L. I., 14-34 Grand Ave Ractliffe, Charles Lionel Brooklyn, 392 Sackman St Berner, Aaron Brooklyn, 1121 Bedford Ave. Dean, Leon W. Brooklyn, 506 Amboy St Epstein, Reuben Brooklyn, 421 Cheater St. Haynes, Nat Brooklyn, 1300 New York Ave Herdman, Raymond C. Brooklyn, 1257 New York Ave Hesse. Henry Richard Brooklyn, 1859 -62nd St Sass, Isidore Brooklyn, 55 Johnson St Skinker, Murray F. Buffalo, 191 Franklin St Bandetaon, Harry Greenport, 829 Main St. Larsen, Carl L. Jamestown, 72 Campbell Ave. Beaumont, William Frederick New York City, 1050 Park Ave Cohn, Ralph I. New York City, 617 Weat 141st St Kahn, Morton B. New York City, 760 West End Ave Littman, Leon New York City, 160 West 100th St. Lopes, Melchor, Jr. New York City, 2740 Marion Ave., Bronx Myers, Tbeobald New York City, 500 Riverside Drive 8icari, Domenic New York City, 404 Riverside Drive Turner, Eugene T., Jr. New York City, 781 East 182nd St., Bronx Wheeler, George D. Richmond Hill, 8512 110th St Hudtwalker, William Theodore Riverhead, P.O. Box 982 Trevor, Bertram Rochester, 55 Hortense St Dwyer, Vincent J. RockyRochester, Point,207 Ave. Cc/o R.C.A.Goldstine,Wiebach, Rattan William E. T. Schenectady, 105 Seward Place Clarke. Varro J. Schenectady, 422 Y.M.C.A. Bldg. Lynn, L. H. Schenectady, 842 Union St Orr, Robert W. Whitestone, L. I Summers, Llewelyn L. B. Ohio Akron, 160 Fir St Ehrisman, Henry 0. Ashtabula, R.D. No. 2 Andrus, Roy E. Cincinnati, 1212 Sassafras St. Verkley, B. M. Cincinnati, 3484 Vine St Wells, Martin M. Columbus, 34 -18th Ave Auckerman, U. H. Columbus, Ohio State E.E. DeptRoetken, A. Allen Dayton, 43 Victor Ave. Franzwa, Frederick J. Gambier, Kenyon College Cottrell. Casper L. Lakewood, 2028 Waterbury Road Hood, William A. Marion, 360 Silver St Ackerman, Francis R. Niles, 424 Allison St De Cola, Rinaldo Youngstown, 3630 Market St Pennock, P. L. Oklahoma Norman, 132 Page St Moffett, Le Roy. Jr. Oklahoma City, 1624 East Park Place Pata, Yaromir J. Tulsa, Radio Station KVOO Golder, Frank E. Oregon Portland, 605 Marguerite Ave. N Trumbull, A. F. Pennsylvania Allentown, 531 N. 7th St. McGee, Michael J. Easton, 528 Centre St Messinger, Reuben R. Lancaster, 36 S. Lime St Russell, Walter L. Narberth, 403 N. Narbertb Ave Bates, Clifford W. Philadelphia, 2030 E. Hazard St Martino. Alphonso E. South Tamaqua. Delp, Paul L. Upper Darby, 297 Springton Road Lewis, Oliver I. Wilkinsburg, 901 South Ave Armstrong, Ralph W. Rhode Island Providence, 160 Cypress St. Adams, Raymond R. Tennessee Memphis, 1961 Harbert St Brooks, Maurice W. Texas Dallas, 2603 Madera St Bennett, Porter T. Dallas, 12051 Elm St Godard, L. G. Dallas, 5004 Goodwin Ave. Melroy, Harry C. Fort Worth, 907 W. T. Waggoner Bldg Zeidlik, William J. Moody McCauley, E. Ray, Jr. Virginia Alexandria, Box 107 R4 Meyers, Paul F. Marion Cummings, G. N. Washington Seattle, 1833 -13th Ave. McAvoy, Edward J. Wisconsin Menasha, 526 Keyes St Peerenboom, Cyril A. Sheboygan, 922 Clara Ave. Flentje, Le Roy G. Stoughton, 313 S. Academy St Turner, Russel S. Wausau, 108 Grand Ave Krueger, Otto J. Canada Montreal, Que., Northern Electric Co., Ltd , Keefer Bldg Cash, J. Allan Montreal, Que., 89 Laurier Ave. E Minorc, J. L. Montreal, Que 4581 Sherbrook St. W. Wilooz, B. B. St. Hyacinthe, Que, 7 Bourassa St Chagnon, Adolphe Applioations for Membership 419

Denmark Copenhagen, Kastelavej No. 3 Bertzow, Johannes Andreas England Chelmsford Essex, 26 Queen's Road O'Neill, R. F. Leigh -on -Sea. Essex, 106 Western Road.. Richardson, F. C. London, N.W.6, 42 Fairhazel Gardens, Hamp- stead Erdman, H. London, SW5, 12 Trebovir Road Megaw, E. London, Berkeley St., Thos. Cook & Son, Ltd Samarasckara, V. R. Parkatone, Dorset, St.Nicholas Castledene Road . White. John Walter Holland 2 Dedelstraat den Haag . Ba rtelink, E. H. B. Japan Kumamoto, Kumamoto Broadcasting Station JOGS Hiroee, M. Kumamoto, Kumamoto Broadcasting Station. Morita, Kszuyoshi Tokyo, Teishin-kanri-renahusho siba Park Tei, Sasaki Kumamoto, ii/o Shimisu-Hosojo Kanayama, Toyoenku Poland Warsaw, ul Krucsa 12 in. 23 Romanowski, Zygmunt For Election to the Junior grade California Oakland, 627 Poirier St Ford, Charles Y. Dist. of ColumbiaWashington, Loomis Radio College Carlson, 0. D. Indiana Valparaiso, 405 E. Monroe St Ratta, Bruce H. Iowa Iowa City, o/o Kappa Eta Kappa, 728 Bowery St Stauffer, Ray Everett Kansas Manhattan, 413 N. 17th St. Kipp, Aaron Michigan Battle Creek, 245 Lake Ave Fay, Lewis C. Montana Bozeman, 201 South Third Pelton, George A. New York Brooklyn, 886 Putnam Ave Osterland, Edmund Brooklyn, 24 Bay 31 St Weinberg, Sidney Ohio Columbus, 75 West 10th Ave Conrad, Willard Oliver Oklahoma Stilwell Neeley, Carl E. Pennsylvania. Philadelphia, 653 North 10th St. Huttenloch, Robert M. Canada Toronto, Ont., 10 Kew Beach Ave. .. Hamox, Harold England Gloucester, 48 Weston Road Carlyon, C. OFFICERS AND BOARD OF DIRECTION, 1929 (Terms expire January 1, 1930, except as otherwise noted) President A. HOYT TAYLOR Vice -President ALEXANDER MEISSNER Treasurer Secretary Editor MELVILLE EASTHAM JOHN M. CLAYTON WALTER G. CADY Managers R. A. HEIBING L. M. HULL L. E. WHITTEMORE J. V. L. HOGAN R. H. MARRIOTT J. H. DELLINGER (Serving until Jan. 1, 1931) R. H. MANSON ARTHUR BATCHELLER (Serving until Jan. 1, 1931) (Serving until Jan. 1, 1932) C. M. JANSKY, JR.

(Serving until Jan. 1, 1932).

Junior Past Presidents RALPH BOWN ALFRED N. GOLDSMITH

Board of Editors, 1929 WALTER G. CADY, Chairman CARL DREHER STUART BALLANTINE G. W. PICKARD RALPH BATCHER L. E. WHITTEMORE W. WILSON

Committees of the Instituteof Radio Engineers, 1929 Committee on Meetings and Papers F. H. KROGER K. S. VAN DYKE, Chairman D. G. LITTLE WILSON AULL WM. H. MURPHY W. R. G. BAKER E. L. NELSON STUART BALLANTINE H. F. °mom R. R. BATCHER, G. W. PICKARD M. C. BATSEL R. H. RANGER ZEH BOIT% B. E. SHACKLEFORD RALPH BOWN N. H. SLAUGHTER H. H. BUTTNER H. M. TURNER W. G. CADY PAUL T. WEEKS L. M. CLEMENT JULIUS WEINBERGER E. T. DICKEY HAROLD A. WHEELER CARL DREHER L. P. WHEELER EDGAR FELIX W. C. WHITE V. M. GRAHAM L. E. WHITTEMORE 0. B. HANSON W. WILSON L. C. F. HORLE R. M. WISE J. W. HORTON IRVING WOLFF L. M. Hinz All chairmen of Meetings and S. M. KINTNER Papers Committees ofInstitute S. S. KIRBY sections, ex officio.

420 Committees of the Institute-(Continued)

Committee on Admissions Committee on Publicity R. A. HEISING, Chairman W. G. H. FINCH, Chairman ARTHUR BATCHELLER H. W. BAUKAT H. F. DART ZEH BOUCK C. P. EDWARDS C. E. BUTTERFIELD C. M. JANSKY, JR. ORRIN E. DUNLAP F. H. KROGER FRED EHLERT A. G. LEE A. H. HALLORAN GEORGE LEWIS R. D. HEINL ALEXANDER MEISSNER LLOYD JACQUET E. R. SHUTE J. F. J. MAHER J. S. SMITH A. H. MORSE A. F. VAN DYCK U. B. Ros All chairmen of Institute sec- J. G. UZMANN tions, ex officio. WILLIS K. WING Committee on Awards Committee on Institute Sections MELVILLE EASTHAM, Chairman E. R. SHUTE, Chairman RALPH BOWN AUSTIN BAILEY W. G. CADY M. BERGER A. HOYT TAYLOR L. A. BRIGGS Committee on Broadcasting D. H. GAGE L. M. HULL, Chairman A. F. MURRAY ARTHUR BATCHELLER All chairmen of Institute sec- CARL DREHER tions, ex officio. PAUL A. GREENE C. W. HORN R. H. MARRIOTT Committee on Standardization E. L. NELSON J. H. DELLINGER, Chairman Committee on Constitution M. C. BATSEL and Laws W. R. BLAIR R. H. MARRIOTT, Chairman C. E. BRIGHAM RALPH BOWN E. L. CHAFFEE E. N. CURTIS T. McL. DAVIS W. G. H. FINCH H. W. DREYER H. E. HALLBORG C. P. EDWARDS J. V. L. HOGAN P. S. EDWARDS G. W. PICKARD S. W. EDWARDS HAROLD ZEAMANS GENERAL FERRIE ALFRED N. GOLDSMITH Committee on Membership 0. B. HANSON I. S. COGGESHALL, Chairman J. V. L. HOGAN F. R. BRICK W. E. HOLLAND H. B. COXHEAD C. B. JOLLIFFE H. C. GAWLER R. S. KRUSE R. S. KRUSE GEORGE LEWIS PENDLETON E. LEHDE R. H. MANSON H. P. MAXIM ALEXANDER MEISSNER A. F. MURRAY C. B. MIRICK M. E. PACKMAN GINO MONTEFINALE J. E. SMITH E. L. NELSON JOHN C. STROEBEL, JR. L. G. PACENT AllsecretariesofInstitute HARADEN PRATT sections, ex officio. H. B. RICHMOND Committee on Nominations C. E. RICKARD MELVILLE EASTHAM, Chairman A. F. Ross ALFRED N. GOLDSMITH W. J. RUBLE DONALD MCNICOL K. B. WARNER R. H. MANSON C. E. WILLIAMS G. W. PICKARD HIDETSUGU YAGI

421

PART II TECHNICAL PAPERS .. : . :, ,-''.- ' ., I. I '...... -.. ...iV:i1111-ar 'e- 11. '1/2 PV : i 1 1;., , A, --,-..6 4., 1_..f...-.! - . ...7.:. j. } .. -=, -A l' -'L _ .1_ Y..5. '',.' .-41:§,dij:-.i' .' . . -ft. .m. _ 1.. .. ,Im, . :I. ' -41--2-;**147r5. ..."' - ' ' r- Pi ,- ,.1,.., = . ,1 .:-2 19..;r? Zql. ? A ...:-I':::=-7,.-- ; " - ..X ..;;... . I%- ' -r...... =-i -,P-ri:41':-.i:'.:-.4i:.--- '-' I.': ri.. -.,,-- _ :...' ..,i z_ Xn.,- F: I.. m ' .-r;1."._r...- -,.'. -t - _ e..: .1.-21. .z. r- t. -2_. !-"i.r_ 0__ ...... : i f-1., '_ r, --t!!",. .=, :','',..:"...4 : 1:::;,;:.: e-:-.,!-.,. 'le*1r.,d7 .r,-:-4 .i ,-i !. '74_1 :41-7': , "rel^r.*, 5.,- rirl'..1,t.-7,, ..i. .7 - 11"lfg--:4 i7.' .0.:: N_,7,f,.-A ,_,7.6; 4L-, , - -.4, ','- ,r,'';: '-,-F-_--:,?- , .t...-..-: .,, ,. :iiijiiiippr ',27-..44 '' . kV2:Tertiel:.',:i46: 7,. vvgi4,A=,,_,- - %.:.-.,!-.;.:,..:7"I'-'_-;,:.-71:117:1-...7.1'.-::t'l:-_-'1.:=Lf::.'lli __27-,JF,,,,....i., ..,..1,1! ..' .:..., 44 ...,":.7,11.,-.f.:,-..7 ; 7,, v:....41.- t.ii....,-7... ,:,.,.. :,..c,.:,,. e.. , ,,,,4'.. ,,,: .. .,,,,,.,..-- i 77, 4._.;--7.-4:-,, 7 2; .....' 1 ; -',-, . - - 7' -.1-: _,.....:.-..4.,,,,, .,, '11-)L11::., ..4 !',:s4r.114.:,.s. '7.'T . :-.'.!-: 'L.' : .. - - L ,i: ,!. ..:..7-ii:i ..1;,.''.E.`--b:ff ,' 'C,I.,tX;;;;',iiit . .. 4"r --x t . ' , '.."--'' ' '-'1 - - , -- . ... '... -. - . - :.'...... ::-7,:,' ',.... l' v...--,_-', :-- 1- 7...-7 'C1 -7t.-' I ii.24 J ' -. .4- `-,-.-- t - Fr -4.; .. --' ' ,' .1''''''''''''.. .. ' ' '' ''' ' - ' .- '." '- 2'"." r ..1,,-:-1 ' ,-i.. -. , . , . - 7r7.'- '...-: '' ;Y',_,,,. .,.t' '..,': ,:_, !...... A.14,:-.,'6.: .. 4:,-7V7 e'i . .. ' . :71--'r .--" ;...... -'''''-'''""7.-,,T.t;'_.,,; ,...... 1-.77.''',,,,1'-'t',,;.''...,7...1";--1.:. L_ ti. ...-,-. ...,1-_,,:,t,!..-:...',.:.--'7 .,-;_f,i1::-.- - .,.,'-' '-', ' ' - ..r, -',,. - .: ,:17 .--1.; -.;-.'''.-,k-., .,. - -. - ii'...P; e. . * I., r.ofi'_,_:%. ,

. - , :. L ---= -%-- ' -.-:...---.7,--r. --;,:' ).. -,.--- ,' ' -"-',--,.4.-e:1---- .J.-4-.)., : - ..'i . - ..)i..'. . 7 :', 1.....4, 7(:,-4;:. ...' ' :....t. 2. ::_ ,,, ..:_ LT - __'-; - ,'..-:'..,.."-;1. /...: -,',". ,i'-',..7..-..,''s,- 1.r. -A ._::-':':.--;','',__.....--', --':.:'.'r.''. ,-'.j7.:-- -c . . :- - .,'_ ., -,..:,.... ,f r.,-.4.,. 4-1+,'..."-' ';',' . -',..-:-.-4'.---1=',_ ' . ,. ., ,,- :2_-_;,,,,,..._::..., ., .. :'.. - ,- , , .. - .- .:!.:"- , ; .. .._ .. ., ..i..- -.--:, - ,t_':.; ...c-!..., ',; i-, --' -.),., 1 7,-,., .- ; ....,-.,'.,';',:::, - t J.- " ....47.' '',.....,;',-,'--.1,'',:', .:71,;','.-,-'-' 7,:. ,.:' .--',;1' ' tit. 11.....,...,-;'..e1",_...,, ..,;-,0. 1,:ij 1 . 7). .., - ....,,,6. ,.,,,..,--.:'1'..t] 'il..' ... :1- - 261-...r.... l''''' ' --',.-1... ..7 i., - i ,..." 'Ifl.:11 T'..j.:51:.1;%''' --,-j:I .- ' "11-.4 ..-. .-.1.1' z..-.%"1- - Proceedings of the Institute of Radio Engineers Volume 17, Number 3 March, 19.e9

RADIO DIRECTION -FINDING BY TRANSMISSION AND RECEPTION

(With Particular Reference to Its Application to Marine Navigation)*

BY R. L. SMITH -ROSE (The National Phyeical Laboratory, Teddington, England)

Summary-This paper presents a critical resume of the performance of apparatus employed for radio direction determination either by transmission or by reception. After an historical summary of results obtained in various parts of the world, a brief description is given of the fundamental principles underlying radio direction -finding. In this section attention is drawn to the application of the principle of reversibility to this art, by the aid of which the behavior of directive radio transmitters can be largely predicted from the more numerous results and greater experience already obtained with directional receivers. The next two sections of the paper give a review of the results obtained in Great Britain during the course of extensive investigations into this subject during the past seven years.Observations obtained from thirteen direction - finding receiving stations, specially erected for the purpose, have been carefully analysed and the performance of the apparatus studied under a variety of conditions, including operation in daylight and darkness, and both oversee and overland. In addition, some two years have been spent in studying the performance of a rotating -loop beacon transmitter, by means of which accurate radio bearing can be obtained with any type of receiving apparatus. The later portions of the paper deal with the application of direction- finding to marine navigation, and of the possible effect of coastal and night errors in connection therewith. The production of night errors on dosed loop receivers by the horizontal component of the electric force in downcoming waves is explained, and a demonstration is given of the manner in which the Adcock aerial system gives freedom from such errors.The paper concludes with a discussion of the relative advantages of direction -finding by transmission and reception for navigation purposes. A bibliography of the subject is appended.

Dewey decimal classification: R125.1. Original manuscript received by the Institute, November 14, 1928. An abstract of this paper was read before the meeting of the Union Radio-Scientifique Internationale held in Brussels, September 5 to 10, 1928. Presented before New York meeting of the Institute, February 6, 1929. 425 426 Smith -Rose: Radio Direction -Finding

1. Historical HE application of the radio direction -finder both as a navigational instrument and as a useful scientific tool in the study of the propagation of electromagnetic waves has been developing rapidly during the past few years. As a result, a number of text -books expressly devoted to the subject of directional wireless have appeared, and the reader must be referred to these't and other published works for an account of the fundamental principles of the art of direction -finding and its historical development. The object of this paper is to present a critical resume of the results of investigations which have been carried out over a period of several years into the performance of direction -finding systems, and the historical background provided in this section will be confined to this aspect of the subject.

(a) The Direction -Finding Receiver The modern radio direction -finder undoubtedly owes its success very largely to the introduction of valve amplifiers, enabling a moderately large reception range to be obtained, and its practical development therefore dates from about 1915. Previous to this, such systems of direction -finding as were in existence were confined to short -distance working and the comparatively crude instruments then in use made accurate systematic observations difficult to obtain. As early as 1908, however, Pickard' observed that large errors might be obtained in the reading of coil direction -finders due to buildings, trees, and other obstacles in the neighborhood. In the diagrammatic representa- tion of his results, the errors are shown to be approaching 90 deg. It was found also by Fessenden' in the years 1901-07, that errors in apparent direction of as much as 20 deg. to 45 deg. might be obtained in the indication of these instruments when receiving over a range of 100 miles. These errors were attributed to a refraction effect resulting from the difference in conductivity of land and sea -water, or even to a varying local conductivity of the ground and of vegetation.In making continuous observations day and night for a week, Fessenden was apparently the first to observe that the errors were greatest during the night, a fact which he attributed to a refraction effect of large clouds of ionized air in the path of the waves.

t See bibliography. Smith -Rose: Radio Direction -Finding 427

A method of obtaining an absolute zero of signal -strength on a small frame -coil direction finder was described by A. H. Taylor', in which the "antenna effect" of the coil is compensated for by a small emf from an auxiliary frame at right angles to the first. Using this system at Washington, it was found that while readings taken in the daytime were fairly accurate those taken near sunset and at night were very erratic. While it appeared that the variations observed on continuous waves of wavelength 13,600 meters were greater than on shorter waves, they were quite serious on damped waves of 1,500 meters wavelength.These variable results were briefly ascribed to reflection and refraction effects occurring during the propagation of the electromagnetic waves over the earth's surface. The liability of the metalwork of a ship to produce a quadrantal error in the readings of a direction -finder mounted thereon was mentioned by Blondel6 in 1919, while the corresponding effects on an aeroplane were later described by Robin- son.' The complete theory of the effect of the metal hull of a ship on a direction -finder was first given by Mesny7 in 1920.Cal- culations from this theory, confirmed by experimental results, showed that the quadrantal error obtained may be as great as 12 deg. This error was shown to be independent of wavelength and to decrease with the height of the frame coil above the deck. Attention was also drawn in the above paper to the approximately analogous case of a direction -finder erected upon a hill or an island, and the resulting quadrantal error which may be obtained in such a case is illustrated by a curve having a maximum value of 15 deg. The phenomenon of the refraction of electromagnetic waves in passing over a surface of suddenly changing conductivity was discussed by T. L. Eckersleys in 1920. Experimental observations made in Cyprus and Egypt on wavelengths between 800 and 1,100 meters showed that wireless waves in crossing a coastal boundary between sea and land might suffer a deviation of as much as 4 deg. This deviation falls to zero for normal incidence of the waves on the coast line, and it was also shown to be negligible for wavelengths exceeding 2,000 meters. In the paper there is quoted an interesting case of a bad day minimum being produced by the reception of two waves from a transmitter, the two waves arriving by different paths and with a phase difference which resulted in a rotating field. 428 Smith -Rose: Radio Direction -Finding

Somewhat similar refraction effects on wireless waves passing from dry to wet ground and across a river were mentioned by Kiebitz2 in connection with experiments on a directional transmitter. The deviations of the waves amounted to 8 deg. or 9 deg. for a wavelength of 550 meters. Some results showing the errors to which a radio direction- finder may be subject due to local conditions were given by Hollingworth and Hoyle." Masses of metalwork, tuned circuits, and overhead wires were found to produce appreciable errors in the readings. In a most valuable paper published in 1920, Round" gave an account, chiefly from his personal experience, of the development and application of the direction -finder during the war.The. manner in which the instrument was perfected as a useful tool for both military and naval purposes was described together with the various types of errors encountered, both by day and by night.Reference was made to the work of Adcock, Eckersley, and Wright in connection with these errors, and a brief indication was given of the means by which they might be eliminated for practical direction -finding purposes. A large amount of experimental work on the intensity and directional properties of the electro-magnetic fieldradiated from an aeroplane transmitter was described by Baldus, Buch- wald, and Hase in 1920,12 while a mathematical treatment of this case was given by Burstyn.'3 The errors in the apparent bearings of an aeroplane at a ground direction -finding station were discussed in detail in their relation to the plane of polarization of the emitted waves. The experiments of Baldus and Buch- wald, in particular, showed that a closed -coil direction -finder on the ground could give errors of as much as 60 deg. in the bearings of aircraft. This fact is interesting in connection with the patent filed by Adcock" in 1919, in which was described a means of eliminating the error of observation of the orientation of aeroplanes. Some further observations on the variable night errors were published in 1920 by Kinsley and Soby." Variations in the apparent bearings ranging up to 50 deg. were recorded on wavelengths between 960 and 17,300 meters, and for ranges of transmission from 40 up to 7,500 miles. Towards the end of 1920, the Department of Scientific and Industrial Research in England, acting on the advice of its Smith -Rose: Radio Direction -Finding 429

Radio Research Board, began the erection of a number of directional radio receiving sets in various parts of the British Isles, attached either to Universities or to Government experimental establishments.The object of these installations was to make regular observations of the apparent radio bearings of various transmitting stations in order to obtain data on the nature, magnitude, and other characteristics of the variations of bearings which were previously known to take place. The general organization and conduct of this investigation was carried out by the author from The National Physical Laboratory, Teddington, as headquarters. Up to the time of the termination of the general investigation in September, 1926, about a quarter of a million observations had been accepted for correlation, these observations covering the range of wavelengths of 300 to 20,000 meters. A detailed record of the work, with discussion of the conclusions drawn therefrom, has formed the subject of a number of official reports published by H. M. Stationery Office, England."In addition to these publications, descriptions of various investigations subsidiary to the main line of research have been published elsewhere. For example, in order to obtain accurate bearings in continuous wave working it was found necessary to prevent any mutual induction between the local oscillator and the receiving frames of a Robinson direction -finder by suitably screening the local oscillator."Such screened oscillators have now been in use for several years inconnection with various direction -finding work, and nowadays they find considerable application in radio measurement work. These experiments were later extended to that of screening assemblies of amplifying and receiving apparatus to prevent the undesired direct induction of signals.Data were obtained by Barfield" for the screening of a complete hut containing the receiving apparatus as, for example, in the Bellini-Tosi direction -finding system. In the course of the same investigation Barfield demonstrated the properties of open wire screens, and these were later applied to the reduction of the error known as "antenna -effect" in radio direction -finding. Various experiments were carried out in the early stages of the main investigation to ascertain the effect of local conditions such as metal -work, overhead wires, trees, etc., on the readings of direction -finders."'"These experiments showed that some quite large errors, ranging up to 22 deg., can be produced by the 430 Smith -Rose: Radio Direction -Finding proximity of such obstacles and emphasize the importance of exercising care in selecting a suitable site for a direction -finding installation. The difficulty in finding any approach to an ideal site was illustrated by the results obtained from the stations selected. In very few cases was the error due to local conditions less than 2 deg.It is fortunate, however, that such a type of error remains constant in value for any particular direction, so that it can be treated as of the nature of a permanent deviation, which can be ascertained periodically from a calibration of the station. In the case of the Aberdeen station, the cause of permanent errors ranging up to 15 deg. was traced to a long iron plate beneath the ground and supporting a sewer duct, over which the direction -finding installation was inadvertently erected.2° During the course of the main investigation each of the three practical types of direction -finder, known as the Bellini-Tosi, Robinson, and single -coil systems, respectively, has been used in some portion of the investigation: and it was naturally desirable to verify whether comparable results could be obtained with any of these systems. A simple consideration of the theory shows that there is no essential difference in the basic principle of any made as the result of some portion of the apparatus rotating about a vertical axis being set in the position in which there is a minimum, or in the ideal case, zero emf induced by the incoming waves. This theory was discussed in a previous publication22 which also contained a description of the results of experiments showing that each system was equally subject to variable night errors. (b) The Direction -Finding Transmitter As an alternative to the use of a special radio receiving instrument for the determination of bearings the directional property can be transferred to the transmitting station, so that its direction can be determined at a distant receiving station by the aid of some characteristic of the emitted radiation. Both of the types of directional transmitting systems which are in use today may be said to date from the time of Hertz, since in his experimental researches Hertz used reflectors to concentrate the radiation from a straight rod aerial and also loops whose transmission or reception properties depend upon the orientation of the loop. Of recent years the properties of reflecting Smith -Bose: Radio Direction -Finding 431 systems used in conjunction with linear aerials have been studied by Marconi? Franklin, and others," and with the application of the beam system to long distance communication considerable development has been taking place in many countries.The beam system has also been adapted as a rotating beacon transmitter? but practical limitations necessitate its operation on the very low wavelengths of 6 to 10 meters. Experimental beacons of this type have been installed at Inchkeith"." and South Foreland," but it is believed that its use among ships has not been very widespread. The predecessor of the closed loop antenna, whether for reception or transmission, is to be found in either the inverted L aerial or in a pair of spaced vertical aerials.The former has developed into the Beverage aerial which is now used as one method of obtaining directional selectivity at the receiving end. A system of inverted L aerials giving directional transmission has been experimented with by Scheller, Buchwald,26 and Kiebitz."For navigation purposes itis also probable that the Telefunken compass" arrangement falls under this heading. This system made use of a series of directional aerials radiating from a central mast, these arials being excited in turn by a rotary switch operating at a speed of one revolution per minute. A non -directional aerial was also provided for the purpose of transmitting a timing signal applicable to the directive system. By noting the interval between this time signal and the reception of the signal from the directional system, the bearing of the transmitter could be estimated from a distant receiver. The combination of the pair of spaced aerials developed first into the Bellini-Tosi system of direction -finding and later into the rotating frame coil.The transmitting equivalent of the Bellini-Tosi direction -finder is to be found in the Radiophare, several of which were in operation on the French coast prior to 1914. In the same manner the single -frame coil can be adapted to transmission since its polar radiation characteristic is the well-known figure -of -eight diagram, and thus the intensity of the radiation varies according to a cosine law with the angle between the plane of the loop and the direction of transmission. One of the great advantages of a loop transmitter is that it can be operated on the wavelengths usually employed in ship and aircraft wireless communication. Various methods of applying the rotating loop transmitter for navigation purposes were 432 Smith -Bose: Radio Direction -Finding described by Erskine -Murray and Robinson in 1922," while more recently Gill and Hecht" have recounted thedevelopment of the rotating beacon by the British Air Ministry and supplied some typical results obtained in the applicationof this system to the navigation of aircraft. The use of two directional transmitting aerials sending complementary Morse signals (such as A and N) upon the same wavelength for the purpose of providing an equi-signal zone along a fixed course appears to have originated withScheller in Ger- many, and to have been investigated by Buchwald" andKiebitz" in 1920. With the substitution of closed loops for the open inverted L aerials previously employed, Engel and Dunmore" published in 1924 an account of experiments designed to illustrate the utility of this system for the navigation of both ships and aircraft. This method is not strictly one of direction -finding since the observer or navigator is not generally able to determine his bearings or position by this means, but only to locate himself along a fixed course as given by the beacon transmitter. For this reason the system is more directly applicable to thenavigation of aircraft flying along fixed routes than to the navigation of ships; and it is from this point of view that considerable development of the method has recently been carried out in the United States of America. This work has been described by Dellinger and Pratt," who have applied a modified form of the radiogoniometer to enable the course or equi-signal zone given by the transmitter to be oriented in any required direction.The limitation of range of the system for accurate direction indications is illustrated by the results of some experiments reported by Pratt" in a separate paper. In common with all closed loop transmitters and receivers for direction -finding this system becomes liable to serious errors at ranges exceeding about 100miles, even whin the receiver is some 2,000 feetabove the ground. For further detailed information on the application of radio communication to aircraft navigation, the reader may be referred to the Bibliography recently published by Jolliffe and Zan- donini." 2. Brief Description of Methods of Direction Determination by Radio Transmission or Reception (a) The Radio Direction -Finder The radio direction -finder is now well known as an instrument which can be used to determine the direction ofarrival of Smith -Rose: Radio Direction -Finding 433 wireless waves, and it will suffice to give the briefest outlin heree of the principles underlying this instrument. The several commercial types of direction -finder now in use employ the same fundamental principle of the reception of vertically polarized wireless waves by a frame coil.In Fig. 1(a), for example, let C represent a plane vertical loop rotating about a vertical axis in the field of an arriving wave whose component electric and magnetic forces are as shown. From the plan view in Fig. 1(b) 7 (a)

(4) Fig. 1 it is evident that the emf induced in the loop by the arriving waves will be proportional to the cosine of the angle a between the direction of the magnetic field and the axis of the coil. The accuracy with which any definite position of the coil may be located depends upon the rate of change of emf with orientation, i.e., the accuracy is proportional to sina. Thus the determination of the direction of arrival of the waves is most accurate when the signal emf induced by the waves is zero. This point is also illustrated by Fig. 2, which shows the theoretical polar reception diagram for a rotating loop. The strength of the signal emf induced in the loop by a wave arriving fromany direction is proportional to the intercept of the vector OA made by the "figure of -eight." The most important feature which requires attention in the design of a practical direction -finder is the avoidance of spurious The term "vertically polarized" is used here to indicate that the electric force of the wave lies in the vertical plane of propagation of the ware. 434 Smith -Bose: Radio Direction -Finding emf's introduced into the system from one or both of the phenomena commonly known as "antenna -effect" and "direct pick-up." The term "antenna -effect" is applied to the property possessed by a frame coil receiver of acting as an untuned vertical aerial as well as a coil for reception purposes.As a result of this, the receiving system may haire induced in it an emf whose phase and magnitude are independent of the orientation of the coil. The signal heard in the telephones will be the sum of that produced by the rotating coil as such and that due to the equivalent aerial effect of the whole receiver. As the coil is rotated it is found that the signal zeros become blurred into broad minima only, and moreover, they may be displaced from their correct positions.The existence of this antenna effect in the system

RE.CEIVINO LOOP. Fig. 2 therefore makes the observed directions incorrect, and also makes the determination of these directions more difficult. Somewhat similar results may be produced by the second of the two causes mentioned above, viz. "direct pick-up."This last term implies that portions of the receiving system, such as the tuning circuits and the amplifier, are having emf's induced in them directly by the incoming waves.These emf's will obviously be independent of the orientation of the main receiving frame, and they will be effective in adding to or subtracting from the signal strength finally heard in the telephones.It must be appreciated that while these stray emf's may be small compared with the main emf picked up by the rotating frame -coil in its maximum position, they become of very great importance when the coil is turned into its minimum position. The methods adopted for overcoming the effects of these spurious emf's are based on the use of somewhat elaborate Smith -Bose: Radio Direction -Finding 435 screening arrangements, with or without the additionof a compensating condenser for the antenna effect.The application of these methods to the practical arrangement of single - coil direction -finders has been described by Kolster," Dunmore," Long,go Mesny,U) and the author;" and the reader must be referred to these descriptions for further details of such apparatus. With the object of enabling the incoming signal to be clearly audible throughout the whole process of taking a bearing, a crossed -coil arrangement was described by Robinson" in 1920. This system was, at one time, particularly favored for use on aircraft, but with some modification described by Bainbridge - Bell," it now finds widespread application to marine navigation. Before the development of valve amplifiers made possible the use of rotatable multi -turn loops, Artom in 1903 and Bellini and Tosi in 1907 suggested and used large frames of a triangular shape, with the ends open at the top apex, for directional wireless communication. The arrangement developed from this and now generally known as the Bellini-Tosi system was fully described in 1908." The large, fixed closed loops employed in this system are connected to a radiogoniometer, an instrument which reproduces in miniature the directive properties of the external field of the waves. Recent developments of this system for use in ships have resulted in the employment of smaller multi -turn loops, which are more conveniently fixed on board. The development and use of the Bellini-Tosi system during the war was described by Round" in 1920, while Horton," Slee," and Mesny" have dealt with some more recent developments. (b) The Rotating Loop Transmitter As an alternative to the direction -finding schemes outlined above, the directional part of the wireless system may be transferred from the receiving to the transmitting end.This is ef- fected in the rotating loop beacon system, which has been developed to a high degree in Great Britain by the Royal Air Force, and which employs a vertical closed loop transmitter arranged to rotate about a vertical axis at a uniform speed of one revolution per minute. The polar radiation diagram of such a loop will be of the same figure -of -eight form as that shown in Fig. 2 for a loop receiver.Thus as the loop rotates the field radiated in any given direction will vary according to a cosine 436 Smith -Bose: Radio Direction -Finding law, passing through successive maximum and minimum values at intervals of fifteen seconds. When the plane of the coil is perpendicular to the geographical meridian a characteristic signal is emitted by the beacon which may be termed the N point. An observer at a distant receiving station upon hearing this signal starts a chronograph.As the beacon rotates the intensity of the received signal varies and will ultimately pass through a minimum or zero value, at which instant it is known that the plane of the transmitting loop is at right angles to the great circle joining transmitter and receiver.If the reading of

Fig. 3-Stop Watch Fitted with Special "Compass -Card" Dial for Use in Taking Bearings trom a Rotating Beacon. the chronograph is observed at this instant of minimum signal intensity it is evident that the bearing of the transmitter from the receiver can be obtained by a simple calculation. To provide for the case in which the observer is due north or south of the beacon, when the N signal would probably be inaudible, another characteristic signal is emitted after a 90 deg. rotation to the corresponding E point.Bearings observed from this signal as a starting point are evidentlysubject to a correction of 90 deg. It is to be noted that since the radiation from the coil is sym- metrical about its plane, a second minimum will be obtained Smith -Rose: Radio Direction -Finding 437 after a rotation of 180 deg. from the first.With the beacon making one revolution per minute, therefore, a line bearing is obtainable in the above manner every half -minute. To fix the position of a receiving station it is necessary to obtain line bearings in this manner from two or more beacons. Since the timing process mentioned above is but an intermediate step in taking a bearing, it is convenient to provide a stop -watch or chronograph used for the purpose, with a dial specially engraved in degrees and points of the compass as illustrated in Fig. 3.If the center second-hand of such a watch is started on the N signal from the beacon, the indication of the hand at the occurrence of the signal minimum will give the true bearing of the observer.Examples of other dials suitable for working from the E point, and on beacons with different times of rotation, have been described elsewhere."(s)

Fig. 4-View of Rotating Beacon inNide Hut at Fort Monckton, Gosport. The development and initial testing of this rotating beacon system by the Royal Air Force have naturally been confined to a study of its advantages over the direction -finding receiver for the navigation of aircraft."During the past few years, however, the author has been privileged to investigate the performance of a rotating beacon station, especially set up for the purpose in England, and particularly to ascertain its advantages 438 Smith -Rose: Radio Direction -Finding and reliability as an aid to marine navigation." A sketch photograph of this beacon is shown in Fig. 4. (c) The Reversibility of Direction -Finding From the brief description of fundamental principles given in the two previous sections it will be gathered that wireless direction -finding can be carried out by rotating a closed loop at either the transmitting or receiving end of a communication link. The directional property is given by the rotating closed loop and the other station can employ any type of wireless transmitter or receiver connected to any type of aerial system. It will probably aid the reader in understanding the performance of each system and their relative merits for any given purpose, if attention is drawn here to the reversibility of the communication system between an open aerial and a closed loop. The reference of the reciprocal theorem to radio communication has been enunciated by Lorentz, Pfrang, and Sommer- feld," who discussed its application to transmission between open aerials, closed loops, or a combination ofthese.The theorem may be briefly stated in the following terms: if an antenna Al transmits to another antenna A2, the signal intensity in A2 is the same as that which would be received in Al if A2 transmitted with the same power and at the same frequency as was previously used by Al. This equality ofsignal intensity in the two cases is independent of the electromagnetic properties of the medium and of the shapes of the antennas. This means that transmissions from a rotating loop beacon being received on an open antenna should produce errorsof the same type as those experienced when the antenna is employed for transmitting to a loop direction -finder.The utility of this theorem in studying the performance of the two systems will become evident in succeeding sections of the paper, and the author has referred elsewhere to the possibilities of investigating local errors in this manner." On theoretical grounds,the reversibility of the Adcock direction -finding system as a means of avoiding night errors has also been established by the present author.72 In the above cases a reservation must be made as to any ir- reversible effects which may be produced by the influence of the earth's magnetic field upon wireless waves travelling through ionized regions of the earth's atmosphere. The possibility of such a departure from the reciprocity relation waspointed out by Appleton in 1925.47 TABLE I (1) (2) Summary of Observations Taken at Newcastle from February 28,1921 to March 4,1922 (3) (4) (5) (6) (7) Transmitting Station Type and missiontrans-Time of Distance bearingTrue Extent Day Observationsof Error of varia-Max.tion No. Extentvaria- of Night Observations Error of varia-Max.tion AranjuesAberdeen S.C.W. 3.3 length (km) 3.5 I (miles)1,034 154 352.0(doge.)184.6 100 65 variation (deige.) 10.5 1.7 (dem)mean+2.0-0.2 (dogs.)- 6.8meanfrom 102 37 (dags.) 17.0t ion 5.6 (dells)mean+1.6+0.1 . mean(dogs.)±-11.1 2.8from BudapestClifdenCleethorpeeChelmsford S.C.W.C.W. 3.84.6 5.8 3.05.8 17.00 I 1,026. 240358355118 256.2147.2158.1111.6 126 95647245 9.33.57.52.25.0 -1.4-0.8-1.3 +- 1.15.21.82.72.8 256527182134 87 4 33.029.512.0 6.04.24.0 -1.1-3.4-1.5-0.9-0.7 +-+18.5 7.23.72.22.6-18.1 KarleborgKarlsborgHorsesColtano S.C.W. 6.0 4.24.52.5 20.0012.0012.15 I 650287948 175.5140.6 62.462.4 -207204141 33 10.2-6.53.16.7 -0.9-1.9-1.4 - + 0.13.41.7+ 3.5 - 128441392 60 23.037.014.5 7.6 +1.7-0.5-0.7-0.1 + 8.05.7+25.9 NantesMoscowKarleborg S.C.W. 4.5 4.22.54.8 21.3008.1521.0008.00 I 1,521 538650 180.4 71.862.4 2718 62 0.32.42.31.3 -2.9-0.4-0.6-1.2 0.0 +- 0.21.36.40.9 1,838 589 34.0 2.07.64.0 +0.4+0.2+0.8-2.3+0.7 ++23.4-12.1 2.41.2- 4.4 ParisNioolaieffNauenParis S.C.W.S. 4.7 2.84.03.23.92.6 10.4009.2008.1511.50 I 1,546 454612454 156.9 96.399.2 1138269202160 50 13.5 3.44.33.05.8 -2.2-0.7-3.1-2.5 -+ 2.91.72.04.46.9 106 -27 12.813.5 - +1.0-2.4 - +- 7.37.2 - PuldhuParis S. 2.83.22.6 09.3021.1011.3012.0011.0019.20 374454 205.6156.9 -365537605 -5.23.03.1 +0.3-0.5-2.3 - +- 1,72.8- 1.9 448172 - 32.5 -8.5 -2.1 - + 4.7-20.6 WarsawSofiaPoldhu I wIrregular transmission. S. 2.03.22.8 21.30 II 1,400 942374 205.6116.8 92.0 6014 7 4.02.81.5 +0.8+0.3-2.2 +±- 0.91.42.2 166 23-17 26.211.5 5.6- +0.2+0.4-4.5-1.0 - +-21.1 3.35.8 I If/ Smith -Bose: Radio Direction -Finding

3. Analysis of the Performance of Wireless Direction -Finders Attention has already been drawn in Section 1 to the carrying out of a large investigation in wireless direction -finding by the Radio Research Board in the British Isles during theyears 1920-1926. In the course of this investigation, thirteen direction- finding stations were used in various parts of the work, which covered a range of wavelengths of from 300 to 20,000 meters. The precise distribution of these stations and the detailed results of the whole investigation will be found in a series of official reports to which reference has already been made."It is considered to be useful here, however, to give a very brief resume of these results with the conclusions drawn therefrom.Apart from the errors due to local conditions at each direction -finding station, which were studied separately and which could be corrected for when necessary, it was found that the apparent bearings of the various transmitting stations observed varied in a very erratic manner under certain conditions. In the work of tabulating and correlating the results it was found convenient to adopt an arbitrary division of the times of observation into day and night periods. The border lines between these periods were taken at one hour after sunrise at the westerly end of the path of transmission and one hour before sunset at the easterly end. (a) General Nature of Variations in Observed Bearings As illustrating the results generally obtained, asummary of the observations recorded at Newcastle on wavelengths between 2,500 and 6,000 meters during the years 1921-23 is reproduced in Tables I and II, while in Table III is reproduceda summary of the results obtained at Orford from 1922 to 1924on wavelengths of 450 and 600 meters.With the mode of separation into day and night periods thus adopted it was found that theex- treme error experienced in the day periods was usually about 4 deg. for any wavelength and distance of transmission, but during the winter months this limit was considerably exceeded on the higher wavelengths, apparently because of an extension of the night conditions until three or four hours after sunrise. As an example of this phenomenon Fig. 5 shows graphs of the daily bearings observed at Slough on the transmissions from the two stations at Leafield and Nantes.The observations were made on the U.R.S.I. signals sent by these stations at 1400 and 1415 G.M.T., Summary of Observations Taken at Newcastle from March 6,1922 to March 31,1923 TABLE II (1) Type (2) Time (3) (4) (5) Day Observations (6) Max. Night Observations (7) Max. Transmitting Station lengthwave- and missiontrans- of Distance bearing True No. variation Extent of (degs.)meanError of varia-meanfromtion No. Extent(dW1.)vans- of (degs.)Errormean of (dogs.)swift.fromtion BerneAberdeenAir Ministry C.W. 4.13.3 (km) 08.30-09.30 21.0017.00 I (miles) 676252154 (dep.)352.0140.8162.5 530247189 (deo.) 10.317.5 5.1 -1.2+0.7 (deo.)+-11.4 6.9 217252111125 37.233.856.030.7tionmean -0.8+0.5+0.8 +17.5+18.0-19.7 ClifdenCleethorpesChelmsfordBudapest S.C.W. 3.44.33.8 5.83.0 09.5008.00 I I 1,026 355118240118 256.2147.2158.1111.6 -118116 857648 12.0 8.5-4.44.01.5 -2.3-1.5-1.6-2.0 - -+ 6.24.94.32.92.4 - 187 706626 21.012.52.57.1 -0.3-1.9-0.6 0.0 ++34.3 1.34.0-11.9 KarlsborgHorsesColeinoClifden S.C.W. 5.05.8 2.54.2 21.5020.12.15 0 0 I 650287948355 175.5140.8258.2 62.4 316171193189 13.6 8.63.32.3 -1.4+0.1-0.8-1.9-0.5 +- 2.20.88.81.43.5 485244594 98- 57.526.015.610.8- +5.1-1.3-1.8-0.4-2.5 - -31.5-±+14.2 7.79.35.4 - KarlsborgNantesMoscowKonigswusterhausen S.C.W. 5.23.9 2.54.84.2 08.00 I 1,521 658538650 180.4100.071.862.4 421501503125 21 10.5 0.65.58.36.2 -1.6-2.6+2.2-0.5 +- 7.14.24.32.8 945260237 9453 32.426.311.939.055.0 -0.8-3.3+3.8+0.1 +28.5-20.7+38.8+15.0 NauenNantesNeaten S. 4.23.22.5 21.3008.1021.0011.5012.15 I 612538 180.4 99.2 1421250 6084 87.0 4.04.55.2 -2.8-2.4-0.2-0.8 -+ 2.43.00.4 1471 93 78.215.7 -1.0-0.2 +10.0- 6.6 OngarNauenParis (GLO)(GLA)(GLB) S.SC.W. 4.74.43.82.9 2.6 09.2008.1519.20 II 238454238 160.2156.9160.299.2 14861239904277252 11.0 5.52.49.57.5 -3.1-2.8-1.4-0.8-0.1-2.5 +-+60.8 6.19.11.45.7- 3.8 432859979321 - 23.934.459.5 -5.8 -2.7-1.7-1.2-0.2-0.8 - --13.0-17.4+30.0-40.4 4 0 - SofiaPoldhuParis 'S.S.S. 3.22.63.22.8 21.5009.5021.5812.0011.30 I 1,400 374454 116.8205.6156.9 114516 42 2.92.41.5 +0.4-1.1-2.7 + 1.3- 1.50.8 163162 37- 48.030.013.2 - -1.9-4.0 - -31.5-21.7- 7.. - Whitehall I =Irregular transmission. C.W. 3.3 09.0021.0017.00 247 164.9 - 43 -6.9 +0.9 - + 4.3 -- 30-37 12.0 -7.1 -0.3+1.8-1.8 - -+ 6.13.8 - 300305OW NI AA Leafield: aal-i A-12.4 km., Transmissions at 1400 G.M.T. V -1 Ai 0011295 p- _avulti. .12,4 )i . Beating ig1 x Indicates signal minimum sharp_..-.e."-,.ea,..-lcAl% 3 October10 17 23 3 November7 14 21 28vki 51924 December 12 19 26 2 January9 16 23 30 6 i.titit..7sayhrfue February 13 20 27 6 13 20 27 3 10 17 24 March April 1 8 15 22 29May 5 12 19 26 3 10 17 24 31 7 IndicatesJune 1925 July signal minimum flat August14 21 28 4 September 11 18 25 2 October9 16 23 30 6 13 20 November December Ak_it 7 4 11 18 True Bearing Nantes: UA, A-9.0 km., Transm ssions at 1415 G.M.T. A 4"..tAfrill'i 180 P.11....1P.fr*".13644544VitYlp3 October10 17 24 3 November7142128512192629162330613202761320273101724191924 152229 5 12 19 26 3 10 17 24 3December January February March April May June 1925 July 7 14 21 28 4August September 11 18 25 2 9 16 23 30 6 13 20 27 4 11 18 October November December Smith -Bose: Radio Direction -Finding 443 respectively, and thus always took place in daylight. The results are plotted as the weekly extreme bearings over a periodof fifteen months from October, 1924 to December, 1925.It is evident from this diagram that during the summer months the observed bearings are steady and fairly accurate, while during

SUNSET AT SOMME AT 100- ONOAA:BRISTOL. BRISTOL 111MMSMITTMO STATION

SICARMIS. km) 80* strit-t7... Okom (A - 70 - 6D= 30'

L51, 100'- ? 00* 4tr go'km Ekm OMOAR (A -4 4 km) M

S,CT AT 9.0.44. Al "f eMMOL NAueA:840TIX. gOt E IftgBLARINCi MAVEN(A -47,k) 8°. 14 70' co' 30 2311, FES. 1200 1400 1600 1800 2000 21100 2100 400010006000800 10001200 24', Fla 1923 HOURS G.M.T. Fig. 6-Graphs of Observed Bearings of Ongar (undamped waves, k = 2.9 and 4.4 km) and Nauen (undamped waves, X =4.7 km) taken at Bristol over a 24 -hour period.

Swart AT &MOOR 7446.41TTIPM51Ancro. BRISTOL AT BRISTOL MO" - 90 TINESEARING 0'40...(A -140k) 84'

ST- Y 46 0.GAR (A- 4.41(m) 9°80: 6 70' ao-

90,5°.- 80' HamH 4.7km) 70°

1 1 1 1 1 60. 1 1 4xkL1200 1400 Iwo18002000 22002400 MO 0400 0600 0600 10001200 20 APRIL 1923 HOURS .M .T. Fig. 7-Graphs of Observed Bearings of Ongar (undamped waves, X = 2.9 and 4.4 km) and Nauen (undamped waves, X =4.7 km) taken at Bristol over a 24 -hour period. the winter months the daily errors assume appreciable proportions, particularly in the case of Leafield. The general nature of the variations can be understood from the graphs given in Figs. 6 and 7, which show the apparent bearings of some fixed transmitting stations as observed every (1) Summary of Observations Taken at Orford from(2) November 13,1922 to March 29,1924 (3) (4) TABLE III (5) (6) Transmitting Station wavelength Typeand Distance bearing True No. Extentvaria-Day Observationsof Error of varia-Max.tion No. Extentvaria- of Night Observations Error of varia-fromMax.tion AntwerpAmsterdam S. meters 450600 (miles)136143 (degs.)114.9 81.1 152 46 (dogs.) tion3.0 (dogs.)mean+0.6 (degs.)meanfrom-1.7 94 9 (degs.)30.0tion1.0 (degs.) mean+1.2 (degs.)+ 0.7 mean CullercoatsCherbourgBoulogneBorkumBerwick S. 600450 285151236220237 95 330.0219.9332.0178.282.3 344519161142 65 4.03.53.02.55.5 -0.2-0.3+0.3+0.2+0.5 +2.8+2.1-1.6+2.3+1.6 294341110127106 45.090.020.015.014.0 +1.0+0.1+0.6+0.9 0.0 +59.1+11.8+12.4-15.8- 7.4 FlushingFlamboroughDunkirkDieppe S. 450600450 123155 9883 334.8116.0154.1188.0 936331294206149 90 4.02.06.02.53.0 +0.7-0.4+0.4-0.3 +2.3-1.4±-3.4-1.7 1.01.5 212173258193 -85 30.045.015.032.0 - +0.6-1.5+0.3+0.9+0.5 - +20.0+24.5-16.6-28.3- 9.4 - MokLisardHelderHavreGrimsby S.B. 450600450 149327148188 246.2200.0328.0 64.065.0 155544977 25 3.04.05.5 -0.3-0.6-0.0-0.5 -2.2-1.6+3.1+2.5 194108563 25.075.020.011.0 44.7+0.2+1.5+0.8 ++58.5+11.2 9.8-13.9 PortlandParkestonParkeetonOstendNorthNiton Foreland Quay S. 450600450 204161 8449.516.5 239.4230.5135.2185.8 401312380 2010-81 7.00.53.04.02.55.5 - -2.8-2.9+0.1-1.2-4.3-1.3 - +0.3-1.6+1.7+2.4-3.2-1.7- 350232404 47-4010 35.024.015.020.0 -4.01.0 -2.9-3.2+0.1-1.2-4.6 - -+13.7+10.4-15.9 2.3- 0.6 - TeddingtonScheveningenRochefort S. 1,800 750500450600000 440117 93 240.6195.8 89.2 129712139110 -72 2.54.55.56.5 - -2.6-0.4-0.5 - +3.8+2.5-3.8-3.9- 498135551106 6852 125.0 35.025.020.013.0 -3.1+1.7+2.8+7.6-0.3-1.1 +12.5+82.7-18.9-13.1+11.4-22.0 Teddington C.W.I.C.W.1,000S.I.C.W. 1,000 1,000 750 9393 240.8240.6 134147 34 2.00.22.51.0 -2.6-3.1-3.0-3.2 ±1.4+0.2+1,5+1.9±0.5-2.0-1.4 221444101284 76 10.020.017.025.0 2.20.5 -3.1-3.0-2.4-2.7-2.7 ++11.8+.9.1+12.1 0.42.2 WilhelmshavenUshantTreguierTeddington St. Gonery S.C.W.C.W.8. 2,600 1,800 1,800 450 293386307 93 231.9224.9240.6240.667.8 304196 65921915 4.05.03.52.5 -0.6+0.2-2.6-1.4-1.7-1.8-2.5 -2.2+2.9-2.3+2.1-1.7-1.3 323548328123120 12 60.035.013.030.0 8.0 -1.9+2.0+3.9-2.1-1.4 +19.1+ 8.5+36.1-+21.5+15.8 6.83.5 Smith -Rose: Radio Direction -Finding 445 few minutes over periods of 24 hours at one observing station. It is to be noticed that the day bearings are much steadier in the summer than in the winter months, but that in either casethe approach of sunset is accompanied by an increase in the magnitude and frequency of the variable errors which continues throughout the night until sunrise.During the bulk of the investigations carried out on wavelengths between 450 and 12,000 meters these variable errors in bearing ranged up to, but very rarely exceeded, 90 deg. as illustrated in the above tables. In some of the later work the observations have been extended into the broadcasting band of wavelengths. A summary of the results obtained is given in Table IV, and this shows that in the

221). 200' 180'

1443°

'Do. 80'

10 29 30 16so Imo 10 20 30 40 20 zoo uo 20 30 40 30 1900 THEO.M.T. Fig. 8-Graphs of Observed Bearings of Bournemouth (undamped waves, =386 m) taken at Ditton Park, Slough. October 20, 1928. case of the observations at Slough taken on thetransmissions from Bournemouth, the maximum error is given as ± 180 deg. This indicates that a single minimum has been followed round the direction -finder scale through this angle and indeed on several occasions a rotation of the apparent bearing through more than 360 deg. has been observed.Instances of this effect are illustrated in Fig. 8. A somewhat similarphenomenon has been previously recorded by Wright and Smith," but on the'_much longer wavelength of 6,000 meters. Another mode of illustrating the variable errors due to night conditions is that adopted in Fig. 9.This diagram refers to (1) (2) Summary of Observations Taken at Slough from May 28 to December 17, 1925 (3) (4) TABLE IV (5) (6) Transmitting Station lengthWave- Distance bearing True No. Extent"ilia'Day Observationstion of meanError of vanes-meanfromtion No. Extentvaria- tion of Night Observations meanError of ariamean -fromtion NewcastleLondonBournemouthBirmingham (meters) 404365386479 (miles) 245 778818 350.2227.6319.1(dege). 83.5 201 76 (dege.) 6.23.0 (dege.)+0.4+2.0 (clogs.)Max.Max.+4.2+1.9 13461378 803537 >360. 105.0(degs.)75.0 7.5 (degs.)+1.3+5.3+3.4 ± 180. +51.4--67.5 4.8(dogs.) (The transmissions were all on telephony of which the carrier wave was treated - so c.w, for reception purposes - - - -1.4 Showing Percentage Variations of Bearings from the Annual Mean TABLE V 1922-23. (1) (2) (3) (4) (5) 1 at Newcastle, (6) (7) (8) Transmitting Station lengthTypewave-and missiontransTime - of observationsnumber of Total Over 2 deg. 'ercentage Variations from Annual Mean Over 5 deg. Over 10 deg. , Over 20 deg. Aberdeen C.W. (km)3.3 08.30 Day 189 Night 125 37.0Day Night63.1 Day6.3 Night28.0 Day0.5 Night13.6 Day0.0 Night 0.0 AirAberdeen Ministry C.W. 4.13.3 21.0017.00 I 247 -- 252111 17.0 -- 82.033.3 0.8 -- 49.5 8.3 0.0 -- 10.1 1.6 0.0 -- 0.0 Berne S.C.W. 3.03.4 I 530 48 217 26 2.13.2 61.315.4 0.00.4 32.7 0.0 0.0 . 10.6 0.0 1.8 CleethorpesChelmsfordBudapest S.C.W. 4.33.8 08.00 I I 116118 76 187 --70 12.930.5 9.2 61.0 --0.0 0.00.02.5 13.9 --0.0 0.0 0.5--0.0 0.0 0.00.0 -- ColtanoClifden C.W.5.8S.C.W.4.3 4.25.8 09.5021.50 I 316169 85 244594 66 0.00.01.0 28.816.4 6.9 0.00.0 9.11.03.4 0.00.0 0.01.2 0.0 0.00.0 KarlsborgBorsea S.C.W.5.0 4.22.5 20.12.0015 I 125171193 485 5398 20.523.8 6.4 50.983.539.8 0.02.6 42.518.910.2 0.0 20.2 7.50.0 0.0 0.01.9 NantesMoscowKonigswusterhausenKarlsborg C.W.S. 4.83.92.55.2 08.00 I 421501503 21 945260237 94 14.211.3 0,00.5 49.640.026.621.9 0.00.6 18.116.9 5.36.6 0.0 0.01.74.28.5 0.0 0.00.20.42.1 Nantes S. 4.2 21.3008.1021.00 84 --03 2.4 35.5 -- 0.0 8.0 1 0.0 1.1 -- 0.0 0.0 NauenNantesOngarNauen C.W.S. 2.94.73.22.5 12.1611.50 I 1421 904250 60 1471 979 -- 19.317.7 2.06.7 50.252.7 -- 0.16.80.0 24.923.9 -- 0.02.2 5.9 -- 0.00.3 0.60.6 -- ParisOngar C.W.4.48.S.C.W. 2.63.8 09.2008.1519.20 I 14861239 277252 432321859 -- 3.00.04.72.1 24.034.7 1.40.0 0.00.40.1 10.614.8 0.00.0 0.00.0 0.02.71.6 0.00.0 0.00.0 PoldhuParis S.8. 2.83.22.6 09.5012.0021.5511.30 114516 42 162 37-- 0.0 65.023.4 -- 0.0 27.0 --2.5 0.00.0 8.10.0-- 0.00.0 5.40.0 -- WhitehallWhitehallSofia C.W.S. 3.33.2 21.0009.0021.5017.00 I --43 163 3037 30.2 ---_ 40.046.045.4 0.0 -- 10.016.5 0.0 J 0.0 -- 0.03.7 0.0 -- 0.00.6 -- 448 Smith -Rose: Radio Direction -Finding records by several observers of the apparent bearings of Karls- borg during its daily transmissions from 2000 to 2030 on a wavelength of 2,500 meters. The observations were summarized in weekly batches and the graphs show the extremes of the bearings for individual weeks over a period of two years.It is very noticeable that the bearings show little variation so long as they are taken before sunset at the receivers, even though

SMOLT AT Ile- IreCt.VMS Omer/yea Moot STATION. 1000 oa - LATIM se- To - Ascsocet 6d xi' re-

6e- BANooc

ee- - 60' 50 - 86,6ToL 40. -

70- 60'- A 40. - acf ee-

-16DO.NOTON 30-

IIII !HI 111111111-1-1-1-1-1-11 f 1 11 1 III 11,11111111111111 .641.1:06 t152030 WITS 4.5 23 122402117 i416M2)93 16j 141 III 96a,m4_v 1921 AaP1 219 3 Fig. 9-Graphs of Weekly Extreme Bearings of Karlsborg (damped waves, X =2.5 km) observed on the transmissions at 2000 G.M.T. during the two years, March, 1921 to March, 1923. darkness has prevailed over a portion of the path of transmission for some time. In spite of the large variations in apparent bearings described and illustrated above, it is noteworthy that during the night periods the observed bearings show no signs of a definite system- Smith -Rose: Radio Direction -Finding 449 atic error, and so the variations are on the wholeequally distributed about the mean value. Another feature to beobserved in connection with the magnitude of the nightvariations is the comparative rarity of the larger errors.A summary of the results obtained at Newcastle during the last year ofoperation is given in Table V, arranged in such a manner as toshow the proportion of the observed bearings which differ byvarious fixed amounts from the mean day bearing. Fromthis mode of expressing the results it will be seen that it is notoften that more than 10 per cent of the nightresults give an error exceeding 10 deg.This fact, combined with the absence of a systematic error at night, is of great importancein the practical application of wireless direction -finding. It thus appears that the chief effects to be observedin radio direction -finding are generally as follows.In the summer time in those latitudes in which the United Kingdom issituated the observed bearings in daylight are fairly constant and,with the exception of the permanent errors already mentioned,they approximate quite accurately to the true geographicalbearing of the transmitting station. Under certainconditions of transmission to be referred to below this accuracy is maintainedat all times and seasons. When these conditions do notprevail, however, the observed bearings show signs of variable errors as the setting of the sun approaches the path oftransmission. These variations then prevail throughout the night, butdecrease to a negligible extent soon after sunrise.As the seasonal conditions are changed gradually from summer to winterthe variations in the daytime increase somewhat; and the nightconditions begin an appreciable time before sunset andcontinue until some time after sunrise. In themidwinter conditions represented by January, the change from day to night effectsis very much more gradual and the differenceis only distinguishable by the difference in magnitudes of the variations. At this timeof year it appears that on the longer wavelengths thevariations may have an amplitude of the order of 10 deg. or moreduring the daylight hours, and this increases rapidly at about onehour before sunset to the night value of 50 deg. or 70 deg.,and so continuing in the most erratic manner until one hour or soafter sunrise. It has now been satisfactorily established thatthese variations in apparent bearing are due to the arrival atthe receiver 450 Smith -Rose: Radio Direction -Finding

of waves from the upper atmospherepolarized with their electric force in a horizontal plane.This explanationwas first put forward by T. L. Eckersley4° in1921 and supported by experimental results. Of recentyears considerable direct confirmation of the theory has been provided,5°,5'"and the study of the propagation of electromagneticwaves through the ionized regions of the atmosphere now forms thesubject of large numbers of papers published in all parts of the world.

(b) Effect of Various Factorsupon Variations in Bearings From time to time throughout theinvestigation the results have been carefully consideredto ascertain if there isany definite relation between the wavelength oftransmission and the nature, frequency, and magnitude of thevariations in apparent bearings experienced. Over the band ofwavelengths from 300 to 20,000 meters no marked difference has beenobserved in the extreme variations recorded, after due allowancehas been made for the distance and the geographicalconditions over the path oftransmission.The outstanding exceptionto this statement is the case of Bournemouth observed at Slough,as mentioned above, in which exceedingly violentvariations have been recorded.It is probable that thisoccurrence is due to a coincidence ofwave length and distance, and the relativemagnitudes of the downcoming and directwaves.The conclusion that, withinwide limits, the wavelength hasno effect upon the amplitude of the variations observedover a period is to be distinguished from the fact that when the wavelengthof transmission is changed instantaneously the error in bearingdue to night effect also changes appreciably. Excellentexamples of this occurrence have been obtained in observingon the transmissions from arc stations, such as Leafield, which employ forsignalling purposesa marking and spacing wave differing in lengthby about 1 per cent. The errors in bearing taken in rapid successionon the marking and spacing waves differed at times by20 deg. to 30 deg. The experience of the authorobtained in Great Britain that the wavelength itself does not havea very marked effect upon the magnitude of the variationswould appear to differ from that of Austin," who states that inthe United States of America the variations on the shorterwavelengths are considerably less than on the longer wavelengths.This conclusion does not, however, appear to be consistent withthe results obtained by

SI TABLE VI (1) Summary of Observations Taken at Teddington(2) during 24 -hour Test from January to (3) (4) (5) (6) (7) (8) June, 1926 Transmitting Station lengthWave- Distance bearing True vationsobser-No. of ExtentWoo- tion of bearingmeanError of varia-fromMax. tion differingPercentage from Meanof Bearings by more than: (9) CarnarvonBordeaux MUU TX 1418(meters) 200940 (miles) 200452 (degrees) 307.3181.8 ' 316139 (dege.)33.0 (degrees) -0.2 -19.1(d mean .1 2 deg.63.3 5 deg.32.6 10 deg. 5.4 20 deg. 0.0 TuokertonSt.RugbyLongKarleborg Amin Island WQK WCILIFTGBRSAQ 161418 800290900460050 3,5213,403 237760 75 287.3330.4287.9146.3 44.4 192416547219147 34.034.519.83.77.47.7 +2.4-3.9-1.3+2.3+2.9+1.3 + 2.0-22.4+27.9+-11.6- 4.63.4 83.785.855.736.9 0.06.4 35.447.711.90.00.0 11.60.09.40.9 0.00.34.80.0 to 452 Smith -Bose: Radio Direction -Finding

Merritt," Bidwell," and Reich" at Cornell University, which show variations in bearings exceeding 100 deg. on wavelengths of 380 to 500 meters.The difficulty of drawing very definite conclusions from observations made on available transmissions from commercial stations within range is illustrated by Table VI, which is intended to show the decrease in magnitude of the bearing variations when the range of transmission is over 3,000 miles. The results given in this table were obtained at Tedding- ton, England, during a series of continuous tests each lasting 24 hours, in which observations were made of the apparent bearings of various European and American stations transmitting on wavelengths between 14,000 and 19,000 meters. These experiments have shown that over similar distances the variations experienced on the wavelengths 2,000 to 9,000 meters are of the same order as those on the higher wavelengths,and also that the actual amplitudes of the variations decrease con- .siderably for a range of transmission exceeding 3,000 miles.In fact, on the American transmitting stations the variations in

OBSERVATIONS OF 04 NOM TOOT COW Q K 011 15-1-46 10 14 -1 -ti.

IF 720' SuNSCI

No5 I I I I t t 1400400 1400 15001000 1101 1001 000 7000 2100 1203 4.500 000 coo woo moo ow ma woo no ammo moo leoImo Tint. e.m.-r Fig. LO-Graph of Observed Bearings of Long Island (WQK, X =16.5 km) taken at Teddington during a 24 -hour period, January 13 to 14, 1926. bearings show no marked distinction between day and night periods.As illustrated in Table VI, observations taken over several 24 -hour periods show a maximum error in bearing of about 3 deg. for the American stations, while on nearer European stations the night error may range up to nearly 30 deg. From Table VI it is seen that of the two American stations observed one gave 94 per cent and the other 100 per centof bearings correct within 2 deg. during the whole series of tests; whereas of the European stations the best example, Carnarvon, gave only 63 per cent of bearings correct to within this limit. The contrast of the observations on the stations at different distances is illustrated by Figs. 10 and 11 which are typical of the results obtained on transmissions from the American and European stations.The error of the daylight readings is in each case attributable to local conditions at the direction -finder. Smith -Rose: Radio Direction -Finding 453

While discussing the effect of distance on the errors in observed bearings it may be mentioned that Pickard" published in 1922 some results, of which a very noticeable feature was that the apparent bearings of European stations observed at Maine, U. S. A., showed much smaller variations than the bearings of some American stations.This difference was attributed to the fact that the transmission from the former was mostly over sea, whereas in the latter case it was entirely over land.It can be shown from theoretical considerations, however, that this is probably not the correct explanation."(d)Mesny" has also found that the bearings on some American stations observed in France show only a very small variation as compared with bearings taken on the nearer European stations. Experiments carried out in Shanghai by Gherzi" during 1923-24 have also shown that while large night errors are observed for distances of transmission

063CIIVNT10.5 Or04HOUR TeST ON V F T at. a-6-26 To a -6 -26 100' 0153' 150' §145* 3 I4 Ia 130' 12.5* SS=abt g ISTI I I III 01 1200 1300 MOO 000 KO ra1900 000 M012100 ammo MOO 01C0 0200 0300 0100 0000 0100 000 ONO (900 1000 1100 1200 TIME O M T Fig. 11-Graphs of Observed Bearings of Ste. Assise (UFT, X=14.3 km) taken at Teddington during a 24 -hour period, June 2 to 3, 1926. of less than 5,000 km (3,000 miles) these errors have an extreme value of about 6 deg. when the range of transmission is from 5,000 to 12,000 km (3,000 to 7,500 miles). The effect of the incidence of sunlight upon the upper regions of the earth's atmosphere can also be demonstrated during a solar eclipse.Fig. 12 shows the results of observations made at Slough on the transmissions from the Manchester broadcasting station during the solar eclipse of June 29, 1927. Although the normal night variations had ceased at about one hour after sunrise, it is seen that the variations were temporarily restored during the period of obscuration of the sun. A discussion of these and other wireless observations made during the eclipse will be found in the official published report." In concluding this section it may be repeated that it is only intended to be a general summary of the results obtained from 454 Smith -Rose: Radio Direction -Finding the co-ordination of a large quantity of data. Anyone who is interested in obtaining more exact information on any portion of the work may be referred to the official reports containing the tabulated results in detail and published by the Radio Research Board. As far as the actual nature and magnitude of the effects observed are concerned, the author believes he is correct in stating that his results obtained in the British Isles are in complete agreement with the observations made in other countries, such as by Mesny and others in France;58 '61 '62by Stoye in Germany;" Austin," Bidwell,'" and Pickard" in America; and by Gherzi" in China.

111111=3161ZaiiiiiNiNIMININN=II

zo FIMMIMINNINNINNIINN

, NM MINI ININNIKINEI MN NI NI MINII NEMRINAIIIIIIIIII IIPPWIIIM...MIPI'MkEWM1111111ElilirliaililINWAZZL....MM. ;.1111=111M=IiiIIMIIIMIr Irdl11111 36 _MU in.= 0.506,woo OM 0620 0530 OSW J4 G.M.T. Fig. 12-Observations of Apparent Bearings of Manchester (2ZY) at Slough during the Solar Eclipse, June 29, 1927. 4. Analysis of the Performance of a Rotating Beacon If the application of the principle of reversibility in direction - finding as stated in Section 2 be accepted, then it is evident that the performance of a rotating loop beacon transmitter can be largely predicted from the results and experience obtained with receiving loop direction -finders.Thus a rotating loop beacon when erected on the same site as a direction -finder will give bearing observations at a distant receiver, which will be subject to the same type of local error and night variations, for example, as the bearings observed on the direction -finder when the distant receiving aerial is used for transmission.Such de- ductions as these have been confirmed by the authorduring the past two or three years in the investigation of the performance of a rotating beacon erected at Fort Monckton, near Gosport, England." Bearings observed on this beacon in certain directions were found to be subject to a small permanent error, which was of an approximately quadrantal nature, and which decreased in Smith -Rose: Radio Direction -Finding 455

magnitude with increase of the working wavelength.46This error was of the same order as that experienced with a wireless direction -finder set up in proximity to the beacon, and was found to be most probably due to some underground power cables. In the choice of sites for future beacons it is evident that a portable direction -finder can be usefully employed in ascertaining the suitability of the site and its liability to produce local errors. In order to ascertain the reliability of this type of rotating beacon as an aid to marine navigation, a number of tests were carried out on ships crossing the English Channel between Southampton and Havre, and Southampton and Jersey. Using the ship's ordinary wireless receiver observations of the bearing of the beacon were made at intervals during each trip and compared with the bearing as given by the Captain of the ship. A typical log of one of these tests is given in Table VII, while the chart in Fig. 13 shows a comparison between the ship's "dead reckoning" and the wireless course in three other tests. As a result of tests conducted on these lines it was found that in the majority of cases the estimated and observed bearings agreed within from 2 deg. to 4 deg., although at times this difference ranged up to 12 deg. Signs of night effects in the shape of indistinct signal minima and wandering bearings were observed at distances exceeding 50 miles, but these were not always coincident with the above differences which sometimes occurred in daylight. In many cases at night and during misty weather, when visibility was very poor, the ship had to be navigated by dead reckoning so that the estimated bearing of the beacon from the ship may be subject to some suspicion. In order to study these night variations in a more satisfactory manner, a number of tests were arranged in which observations were made continuously over a period of twelve hours or more from various fixed positions.The places of observation were selected so as to provide a variety of ranges and also to show the difference between transmission over sea and over land.Some of the observations were carried out in ships moored in dock or alongside a light vessel, while other observations with transmission entirely overland were carried out at Slough and Ted- dington. An analysis of the readings obtained inthis manner in daylight shows that for ranges up to 56 miles the maximum departure of the observed bearings from their mean or correct 456 Smith -Rose: Radio Direction -Finding value was 3 deg., while up to the maximum range of 119 miles the greatest deviation was 6 deg. In most cases over 90 per cent of the observations were correct within 2 deg. When the observations were carried out at night, the above accuracies were maintained for distances overland of 14 miles and oversea of 23 miles. At distances of 92 miles entirely oversea the night errors experienced ranged up to a maximum value of 18 deg., but it is to be noted that in these cases over 84 per cent of the bearings were cor- TABLE VII Summary of Observations Obtained on a Return Voyage from Southampton to Havre, September 22 to 24,1926. Continuous Wave Transmission, X -525 in

From information supplied by Captain of ship Wire - 141 Date G.M.T. ' From Fort bear- Remarks Position and Monekton ing how obtained from Die- Bear- beacon Lance ing 22.9.26 2240 Southampton.A. miles degs. degreesMean of several 14.3 127 126 readings. 23.9.26 0102 Beyond Nab.V.B. 17.5 324 323 0132 D.R. 26.0 328 329 0202 D.R. 35. 330 331 0232 D.R. 43. 331 331 0302 D.R. 51. 332 332 0332 D.R. 60. 333 334 0532 D.R. 85. 335 335 0602 D.R. 94. 334 3351 Night effect 0632 D.R. 96. 330 331J observed 0722 Havre. A. 105. 328 330 1702 Havre A. 105. 328 329 Mean of several readings. 24.9.26 0012 D.R. 95. 330 331 0032 D.R. 88. 330 3291 Night effect 0102 D.R. 79. 329 327 f observed. 0132 D.R. 69. 329 325 Night effect 0202 D.R. 60. 329 325 observed. 0232 D.R. 52. 329 325 0402 D.R. 24. 329 326 0432 V.B. 15. 325 326 0442 2 miles from Nab. V.B. 11.5 324 323

Notes.- (1) A =moored in dock, V.B. =visual bearings, D.R = dead reckoning.(2) Night effect observed at 60 miles and over. (3)Maximum difference between estimated and observed bearings =4 deg. at night.(4) Sea moderate throughout. rect to within 5 deg. When the oversea distance exceeded 100 miles the errors became more serious. For overland transmission the errors were much larger, and at a distance of 55 miles errors up to 32 deg. were experienced, while only 66 per cent of the observed bearings were correct to within 5 deg.In spite of such large errors, however, it was found that the mean bearing of the beacon from any position was practically the same by day or night. This implies that the systematic error of the night variations is very small, so that the effect of errors of individual ob-

458 Smith -Rose: Radio Direction -Finding

At each site a simple receiver with a vertical aerial was employed for the reception of the signals from the rotating beacon. Adjacent to this, but at a sufficient distance to avoid mutual interference, a single -frame coil direction -finder was set up for the purpose of taking ordinary direction -finding bearings while the beacon loop was fixed with its plane approximately in the direction of the receiver.As a further precaution against producing a spurious reading on the direction -finding set the vertical aerial was disconnected during the taking of bearings on the direction -finder. A few tests carried out on land showed that apart from the occurrence of local errors due to the site selected, the two systems gave observed bearings of similar accuracy. In order to study the variations in bearings experienced with the two systems over extended periods, further experiments were carried out on a ship at Jersey and on land at Slough. The analysis of a large number of readings taken in this manner confirmed the similarity in accuracy of the two systems, and showed also that results obtained with continuous wave transmission were similar to those obtained with interrupted continuous waves. At night time the variations in observed bearings were quite serious in all cases.The extent of the variation ranged from 20 deg. to 57 deg. in the case of the rotating beacon bearings, and from 69 deg. to 130 deg. in the case of the direction -finding bearings. Under these conditions, however, over 86 per cent of the bearings taken on the rotating beacon were in error by less than 10 deg. while in the case of the direction -finding bearings over 62 per cent were in error by less than 10 deg.As already mentioned it is now well-known that these errors are due to the reception of waves from the transmitter deflected in the upper atmosphere, and it is realized that in the results quoted above the variations on the direction -finder are exaggerated by the fact that the loop transmitter radiates very much more to the upper atmosphere than is the case with a vertical aerial.

COMPARISON TESTS AT SEA Some of the test runs made on ships between Southampton and Jersey, discussed above, were carried out in a ship which is fitted with a Marconi direction -finder using fixed frame coils on the Bellini-Tosi system; and the opportunity was thus provided of comparing the two systems of obtaining wireless bearings Smith -Rose: Radio Direction -Finding 459

under actual sea -going conditions. The results obtained on one trip, in which observations were made successively on the rotating beacon and on the direction -finder with the beacon fixed in the maximum signal position, are given in Table VIII. Considering first the results in Table VIII it is seen that with two exceptions the direction -finding bearings agree with TABLE VIII Summary of Observations Obtained of a Return Trip between Southampton and Jersey. September 14 to 16,1927

From information supplied Wireless by Captain of ship Bearings: Type of Re - Date G.M.T. From Fort marks Position and Moncktonfrom by mie- how obtained bea-D.F. sion Die-Bear-con tam* ing milesdegedegs.degs. 14.9.27 2340 Off Egypt Point. V.B. 9. 90. 92. 88. I.C.W. 15.9.27 0010 Off Yarmouth Pier. V.B. 16. 78. 74. 75. C.W. 0040 nr. Needles V.B. 22. 68. 65. 67. I.C.W. ) Sea 0610 At Guernsey A. 110. 34. 36. 33. C.W. 0710 St. Martin's Pt. D.R. 111. 31. 34. 33. C.W. rough 0740 nr. Grosnez Pt. V.B. 114. 27. 30. 37. I.C.W. 1540 At Jersey A. 116. 21. 22. 26. I.C.W. Means 1610 At Jersey A. 116. 21. 22. 24. C.W. of sev-

1640 At Jersey A. 116. 21. 22. 25. I.C.W. eral read- ince 16.9.270710 nr. LaCorbiere. V.B. 119. 25. 28. Sigs.C.W. too weak 0940 nr. Casquets V.B. 98. 35. 39. 36. I.C.W. 1010 beyond Caw:meta V.B. 88. 39. 36. 36. C.W. 1040 49 54' N. D.R. 78. 40. 38. 42. I.C.W. 2- 17' W. Bea 1110 50° 2' N. D.R. 68. 41. 38. 41. C.W. 2° 9.5' W. 1140 50° 9.5' N. D.R. 55. 43. 39.41. I.C.W. smooth 2° 2.5' W. 1310 Outside Needles V.B. 30. 52. 48. 57. C.W. 1340 Inside Needles V.B. 20. 71. 70. 71. I.C.W. 1410 Lepe Middle V.B. 14. 82. 82. 93. C.W. 1440 Beacon Buoy. V.B. 10.119.121.119. I.C.W.

Nnte..-(1) A =moored in dock. V.B. =visual bearing. D.R =dead reckoning.(2) Maxi- mum difference between estimated and observed bearings on beacon =4 deg. in daylight. The observed bearings have been corrected for land errors in accordance with the previous calibration of the beacon in the area occupied by the ship.(3) Maximum difference between estimated and D.F. bearings =11 deg. in daylight.Part of this difference is probably due to land effects for which a correction is not readily applicable. the estimated bearings to within 5 deg. Of the two exceptions, one shows an error of 10 deg. at a range of 114 miles, while the second shows an error of 11 deg. at 14 miles distance, when the ship was proceeding up the Solent.It is thought that a portion of these errors, but probably not more than 3 deg. or 4 deg., is due to a land deviation effect which is difficult to compensate for 460 Smith -Rose: Radio Direction -Finding on a ship direction -finding set. During the same run, the rotating beacon bearings, after correction for the 1 deg. or 2 deg. of error in some positions from the calibration curve, show a maximum departure from the estimated bearing of 4 deg. In searching for an explanation of the direction -finding errors above, it is to be remembered that the direction -finding bearing is observed relative to the ship's head and that, therefore, the accuracy of the bearing is limited to the accuracy with which the ship's compass indicates the instantaneous direction of the ship's head. On the ship in question the direction -finding set is operated from the wireless room and on a signal being given to the bridge the compass reading is taken by one of the ship's officers. In anything but a very calm sea it is naturally difficult to keep the ship's head steady to within one or two degrees, and in a rough sea the swing may amount to five degrees or more. Also there is probably a lag between the actual direction of the ship's head and the reading of the compass.Finally, unless it is calibrated at frequent intervals, it is doubtful if the reading of a magnetic compass is accurate to within 2 deg. or 3 deg. over all parts of the scale.These considerations may result in the accuracy of a ship direction -finder installation being appreciably inferior to that of a similar apparatus on land. The errors due to the above cause will probably decrease with an increase in the size of the ship, but the ship mentioned above is probably typical in size to many which will, in the future, utilize wireless bearings for navigation. It is also to be remarked that the ship direction - finding set is subject to a quadrantal error, for which correction or compensation is necessary. On the other hand the rotating beacon method of obtaining wireless bearings is free from all these objections.Except for the interfering effects of noise and general vibration, the accuracy of the observed bearings is the same whether the ship is in dock or at sea.

RECEPTION OF BEACON TRANSMISSIONS ON LOOP AND AERIAL Since it may sometimes be required to receive and observe signals from a rotating beacon on a closed loop or direction - finding set instead of on an open aerial, it was considered to be useful to make a brief comparison between the two cases. When working in daylight or under such conditions at night as to be Smith -Rose: Radio Direction -Finding 461 free from night variations, it has been found that the bearings observed on a loop agree with a maximum departure of one degree, with those observed on an aerial. When operating under conditions of night variations, however, the errors encountered on the loop receiver are greater than those observed on the aerial. This effect is due to the fact that with downcoming waves arriving at the earth's surface the resultant horizontal magnetic field (operating on the loop) is of greater intensity than the vertical electric field (operating on the vertical aerial).".52These results have been confirmed for both continuous wave and inter-. rupted continuous wave transmission.

NIGHT ERRORS EXPERIENCED UNDER STRICTLY REVERSIBLE CONDITIONS Reference has been made in one or two places in this paper to the application of the reciprocal theorem to wireless direction - finding.While the form of comparison of the rotating beacon and the direction -finder adopted in the tests described above is the correct one for determining the relative advantages of the two systems for marine navigation purposes, it does not constitute a strict reversibility of the system, aerial to loop. By erecting a direction -finder in close proximity to the rotating beacon and using a distant aerial either to transmit to the direction -finder or to receive from the beacon this reversibility is achieved.In this manner it was shown that both local and night errors are of the same order on the two systems and a graph of bearing taken at night under these conditions is shown in Fig. 14. 5. Conditions Affecting Direction -Finding as an Aid to Navigation In the application of a direction -finding system to either marine or aerial navigation it is important to understand clearly the exact conditions under which the observed bearings are accurate, and in other conditions to appreciate the order of magnitude of the possible errors involved and the means of mitigating these where possible.It is generally conceded nowadays that an accuracy of 2 deg. in observed bearings is suitable for navigation purposes, particularly when it is borne in mind that the observations may be repeated at frequent intervals. An important factor to be noted from a navigational point of view is that of any effect of fog on direction -finding, since it 2015 Observations on Transmissions from Fort Monckton Rotating Beacon made at N. P. L, Teddington. ri-g.a IP. I A° Etioio' 5 0.05.175'0,11a 0ig CE 0 4 iii 1! 5 1 I I wzb2v' ig. W8 .....9 0 . 10 15 CI I 111111 I I co cp ii IF .-...13,,oo aii-DO 0. , 20 0030 0040 0050 0100 II 0110 0120 0130 II ill Hours 0140 IllG. M. T. 0150 0200 0210 0220 0230 0240 0250 0300 SOCUa0 .4 0a. V .rf CC5CD o 1015 Observations on Transmissions from N. P. L, Teddington made i on D. F. Set at Fort Monckton ta. 0 .,.° cieto*i 0e'g ps...a cpo 5 CA01 pt.3>41II CP4 to<4. 1 0 cii.1coCD EL 0El 2 0 10 c 15 oa 20 0030 0040 0050 0100 0110 0120 0130 Hours 0140 G M. T 0150 0200 0210 0220 0230 0240 0250 0300 Smith -Rose: Radio Direction -Finding 463 is chiefly during foggy weather that the majority of direction - finding stations are called into action. On several occasions the author has taken particular notice of the existence of fog at times when observations were in progress with a negative result. On one occasion in particular the fog was spread over the British Isle and a large portion of Western Europe, but the directional variations showed nothing beyond the usual day and night effects.These results form a confirmation to the observations of Rothe," who concluded that no variations in ordinary atmospheric conditions would account for the small variations in bearings observed in the daytime. During the course of the many experiments carried out with the direction -finder and the rotating beacon it has been found that the type of wave transmission employed has no effect upon the bearings observed, and that the errors and effects encountered are of the same order for damped, undamped, and interrupted undamped waves.®",44 (a) Errors Due to Local Conditions In Sections 2 and 3 above it has been mentioned that various conditions may be present in the neighborhood of either a direction -finding station or a rotating beacon to cause an error in the bearing observed on a distant station.From various detailed investigations into the causes of such errors it is known that the most prominent effects are due to masses of metalwork and wires, either above or below the earth's surface, and to trees.","In most cases where the directional station is situated on land it is possible to select a site which is largely if not entirely immune from such effects, and provided that the residual errors do not exceed 2 deg. or 3 deg. in magnitude they can be definitely ascertained by a calibration at short distances, and a correction obtained for application to other observed bearings.In the case of the use of a direction -finder on board ship it is impossible to be clear of the metalwork of the ship itself and the resulting bearings are subject to a quadrantal error, which can either be compensated by circuit adjustment. or corrected for from a chart."'"'" (b) Coastal Errors Another type of error (which might be classified as one due to local conditions, but which, on account of its importance in 464 Smith -Rose: Radio Direction -Finding the application of direction -finding to marine navigation deserves a special section) is that due to the deviation of wireless waves in crossing a coast -line, when the path of the waves lies approximately parallel to the coast.In the experimental work carried out at Orford, Suffolk, the coastal error on wavelengths of 450 and 600 meters was found to be of the order of 3 deg. or 4 deg. when the direction of transmission was within 20 deg. of the coast -line. In one instance, in which the wavelength was syste- matically increased from 500 to 2,600 meters, the corresponding error decreased from 3.2 deg. to 1.4 deg. On higher wavelengths the observed coastal error was less than one degree.In every instance the error was such as to indicate a bending of the waves towards the normal to the coast -line, in passing from the sea to the land side of the boundary.These experimental results are in complete agreement with those previously obtained by Eckersley8 but, as already pointed out," they are inconsistent with the explanation of the deviation as a coastal refraction effect due to the difference in the superficial velocity of wireless waves over land and sea.For the theoretical aspect of the problem would appear to indicate a deviation much smaller than that actually observed, and also in the opposite direction. More recently some experiments on the propagation of waves across a coastal boundary have been made by Bailinler and Zenneck." Their results generally confirm those given above as to the magnitude and direction of the deviation of the waves in passing from sea to land. They further showed that the deviation at the coast is unaltered by the change in sea -level from high to low tide. Worlledge" has also given brief details of some experiments carried out at a direction -finding station in which indications were given of the reflection of wireless waves from the landward side of a coastal boundary. As a result of this phenomenon errors in observed bearing and indistinct minima due to interference may be produced at a direction -finding station when observing on transmissions emanating from a station on the same side of the boundary. In connection with these coastal errors it is worthy of note that, with the accumulation of experience of the use of direction - finders on board ship, it is now becoming customary to mark out on charts the "arcs of good bearings" cf various transmitting stations, within which the results of observations are found to be reliable."Similar arcs can, of course, be provided for shore Smith -Bose: Eadio Direction -Finding 465 direction -finding stations or rotating beacons; and their bounda- ries would be found in the initial calibration of the station for the permanent local errors. (c) Night Errors and Their Elimination As the result of the analysis of a large quantity of data obtained with wireless direction -finders it can be stated that the minimum distance at which night variations have been con- sistently 'experienced is about 30 miles for overland working, observations taken at closer ranges than this showing a maximum error of only 2 deg. or 3 deg.Mesny58 has stated that night variations may occur for ranges of transmission as small as 15 miles, but according to the author's experience night errors are small at such distances.For example, observations of the bearings of the London broadcasting station taken at Slough, 18 miles away, have shown a maximum error of less than 5 deg. during many months (see Table IV). Similarly, the bearings of the Air Ministry in London, at Teddington, distance 17 miles, have shown a maximum error of 4.5 deg. while 99 per cent of the readings were within 2 deg. of the mean bearing. On the other hand, Ongar, which is 30 miles away, gave an error up to 8 deg. at the same observing station, and only 14 per cent of the readings differed from the mean by less than 2 deg. When, however, the path of transmission is entirely oversea, as was the case in the observations carried out at Orford on the transmissions from various ships, the above minimum distance is increased to about 80 or 90 miles."This fact is of great importance in the application of direction -finding to marine navigation, and it is perhaps fortunate that the usual conditions connected therewith are that the path of transmission is entirely oversea and that the observed bearings should be accurate at distances which are usually less than 50 miles and never greater than 100 miles.It is now known that under such conditions direCtion-finding is sufficiently accurate for navigational purposes. The explanation of the greater range freedom from directional errors due to the diminished attenuation of the direct wave travelling oversea was first given by Wright" in 1920.An exactly similar result has been obtained from the experiments carried out at sea in the observation of bearings from a rotating beacon." Even in the presence of night errors their effect can be 466 Smith -Bose: Radio Direction -Finding largely mitigated with either system of direction -finding by taking a series of observations in succession over a period of a few minutes, when the mean bearing so obtained will be much more accurate than a single reading. As already mentioned in a previous section it is now well known that the night errors experienced in the use of closed -

al' et ;e3-e4 -0 e,- e2 e3- e4-0 :. No EMF E, IN LOOP

PLAN PL N

I. HORINNTAU-Y PROPAGATED vERT1CALLy DOWEECOMItie VERTICALLY POLARISED POLARISED WAVE. BEARING CORRECT. WAVE. BEARING CORRECT. e- ea - 0 e.- et- e e4- 0 (e3 -e4) IS pnA,snurn

PLAN PLAN 117--W".

LLL. DOWNCOMING HORNOMEALLY POLARISEQ Doweacmot0moerurrreuir DOLAROSED WAVE. Low el MAAIROM OMAN. POSITion WAVE. LOOP IN MINIMUM. Meow tr. auitoth-90.

MS 4

,c.'

PLAN PLAN

yy.Two w,,m311. VERTICALLY POMPON Y RECEPTIONOF DIRECT AND DOVINCONINO WAVES. ERROR IN BEARING ..C C. ERROR IN We" c. 'WM"

Fig. 15 coil direction -finders are due to the action of the horizontally polarized component of the downcoming waves. The effect of the horizontal component of the electric force in producing an error in observed bearing is illustrated by the series of diagrams Smith -Rose: Radio Direction -Finding 467 forming Fig. 15.It is generally assumed that the downcoming waves have travelled through the upper regions of the earth's atmosphere without deviating laterally from the great -circle plane through the transmitter and receiver. The author believes he is correct in stating that this assumption remains undisputed at the present time, except for wavelengths below about 50 meters.It will be evident, therefore, that any receiving system which is unaffected by a horizontal component of electric force would be free from night errors, even though the vertically polarized downcoming waves might still produce variations in received signal strength. A direction -finding receiving arrangement which fulfils this condition was patented by Adcock'* in

,..OPPOSED AERIAL -PAIR. Ar- se.

PIV T.

110 APPARATL$ yepTIcaL IN xi SCR ENED

Fig. 16 1919, but this system does not appear to have received practical consideration until Mr. Barfield and the author'* experimented with it in 1926. The simplest form of the Adcock aerial system is a pair of spaced vertical aerials arranged to rotate about a central vertical axis, as shown in Fig. 16, thus forming the equivalent of the single closed -coil direction -finder. By making all connections to the centers of the aerials the horizontal members of the system are compensated so that no emf is induced in the system by a horizontal electric force. To obtain increased sensitivity larger aerials set at a greater spacing may be employed, and when the system becomes too large to rotate, two pairs of such aerials may be employed in conjunction with a radiogoniometer exactly as employed with closed loops in the Bellini-Tosi direction -finder. The schematic arrangement of such a system is shown in Fig. 17, while the photograph in Fig. 18 gives a view of the practical arrangement employed in a series of experiments which have been described in detail elsewhere.'* The type of result obtained is illustrated by the curves given in Fig. 19, which shows the 463 Smith -Rose: Radio Direction -Finding results of simultaneous observations made on an ordinary closed- coil direction -finder and on one of the Adcock type on the transmissions from a British broadcasting station.It will be seen from this diagram that while the closed -coil set during this period of observations was giving very bad night variations, the corresponding errors observed on the Adcock direction -finder were negligible.It is to be concluded from these and similar results obtained on the transmissions from other broadcasting stations that lateral deviation plays a negligible part in producing the large and variable errors which are obtained at night on the present type of closed -coil direction -finding set, and that therefore these errors are almost entirely caused by the arrival of downcoming waves polarized with the electric force horizontal. INSULATING ROPES UP

II II

Tklx.compm,11 PoLes."Ir Itt\

II it-- t- II ", (1-''`.1A-

' "GROUND LEVEL. Fig. 17 It will be appreciated that it follows from the above conclusions that such a system may be used as a direction -finder which gives the true great -circle plane of arrival of wireless waves, whatever may be their state of polarization or their angle of incidence at the earth's surface.The system should, therefore, have important applications as an accurate direction- finder under night conditions, or for use in observing on transmissions from aircraft at high angles of elevation, when the ordinary closed loop type of instrument is subject to large errors. Conversely the Adcock aerial system may be used in place of the closed loop at a directional transmitting station for use either as Smith -Rose: Radio Direction -Finding 469

a course -setter or as a bearing indicator, with freedom from the night errors which have already been experienced.u."The practical development of the system both as a transmitter and as a receiver towards the above ends is now being pursued in Great Britain.

Fig. 18-View of Hut Suspended at the Center of an Adcock Aerial Sys- tem.The hut contains the radiogoniometer and receiver used in taking bearings free from night errors. 6. Relative Advantages of Direction -Finding by Transmission and Reception for Navigation Purposes (a) Location of the Direction -Finder At the outset it must be admitted frankly that, owing to the more advanced state of its development at the present time, the wireless direction -finder when used in a fixed position on land gives a somewhat superior accuracy to the rotating beacon; for it is not easy to obtain bearingson the rotating beacon to a greater accuracy than two degrees, whereas a good land direction - finding station should give bearings reliable to one degree. This, 470 Smith -Rose: Radio Direction -Finding however, is not the whole story and reference must be made to the conclusions of a controversy of long standing on the matters of the location of the direction -finder, whether on ship or shore, and of the individual upon whom it is most desirable to place the responsibility of observing the bearings.Without giving the full details of this controversy it may be stated that it is very desirable for the navigator to take the wireless bearings personally or have these observations made by somebody under his direct control and supervision.Since the navigator is responsible for the safety of his ship it is unreasonable to expect

1 1 SUNSET 1 15 5 O.M.T. QeserevATsones sy 51NOLE-COR.O.F.sysimm SHOWN THUS 0-0-0 320 Aococx's

?130 1111 Hill

11111..

S rtri too grill 0 z

N ao

ao 6 III o co 80

1810 IS 25' 30' 30' 53 1900 05 10' TIME, G.M.T. Fig. 19-Observations of Bearings on Bournemouth, December 10, 1925 (X =386 m) him to rely, particularly in emergencies, upon observations made at a shore direction -finding station by a man quite unknown to him, and of whose skill and reliability he is quite unable to judge.In addition there is the possibility of error in signalling the bearings from shore to the .ship, the interference to other wireless traffic in doing so, and the point of some strategic importance that the information as to the ship's bearings or position is broadcast to all wireless stations in the vicinity.It will be evident, therefore, that the most desirable position for the wire- Smith -Rose: Radio Direction -Finding 471 less direction -finder is on the ship itself, and this view is confirmed by the fact that during the past few years an increasing number of ships have been fitted with direction -finders whereas the number of shore direction -finding stations has never been large and is decreasing. The true comparison to be made, therefore, is between the rotating beacon transmitter set up on shore and the direction -finder installed on board ship.In the following paragraphs the discussion of this comparison is classified according to the superiority which one system appears to have over the other.

(b) Superiority of the Ship Direction -Finder over the Rotating Beacon It was considered at one time that with the installation of the direction -finder on board ship the necessity for special transmissions for wireless bearing purposes would be eliminated. This, however, has been found not to be the case and it is now the practice to erect fixed transmitting stations or beaconsu for the sole purpose of providing special transmissions for the use of ship direction -finding sets. These beacons are usually located near lighthouses or on light -vessels at points of importance for navigation.The use of the direction -finder is not limited to such beacons, for bearings may be taken upon the transmissions from any land station the position of which is accurately known. It has been found, for example, that some broadcasting stations form very useful fixed beacons operating for many hours of the day when it is desired to take wireless bearings on a ship at sea. The navigator with a direction -finder under his control thus has the opportunity of taking a series of cross-check bearings and so of improving the accuracy of determination of his position. A further advantage of the ship fitted with the direction- finding set is that the opportunity is provided for observing bearings upon the transmissions from other ships. This may be useful when contact is made with a ship which knows its own position, but it is of greater importance when applied to the location of ships in distress, as has already been demonstrated in several practical cases. Moreover, there is the possibility that the future development of the ship direction -finder will provide a means for reducing or avoiding collisions between ships in foggy weather. 472 Smith -Rose: Radio Direction -Finding

These advantages are not possessed by the rotating beacon system which provides the ships with the opportunities of obtaining bearings only during the operation of such beacons, and in those areas of the world in which these beacons are installed. The navigator with a direction -finding set under his immediate control would probably use it very frequently and would thus soon become aware of the occurrence of any fault or error. With the rotating beacon system he would be dependent upon information from the shore station staff as to any such errors; but against this it may be urged that such a staff would be moreex- pert at dealing with and removing the error than would the average navigator or wireless operator.

(c) Superiority of the Rotating Beacon over the Ship Direction- Finder By way of countering the remarks made in the above, section it may be said that the rotating beacon transmitter retains all the advantages which were sacrificed when the direction -finding installation was moved from the shore to the ship.These advantages comprise the fact that the directional part of the system is fixed in position on solid ground and that it can, therefore, be accurately oriented and calibrated at installation, and also that it can be under the continuous supervision of experts.One directional shore station will provide service to an almost un- limited number of ships, whereas each ship direction -finder must be carefully installed and operated by experienced men. One of the great adiantages of the rotating beacon system is that no apparatus beyond a simple wireless receiver and a chronograph is necessary on the ship itself; a direction -finding service is thus provided for ships of all classes from the largest to the smallest. In the case of many small ships it would not be practicable to install a direction -finder with any pretense to accuracy; and to the large ship already fitted with a direction - finding set a rotating beacon would be an additional asset in enabling further bearings to be taken either for check purposes or for position determination.Furthermore, the method of observing bearings is so simple that they can easily be determined by the navigator himself in any portion of the ship without necessarily equiring the assistance of a wireless operator. It has already been hinted that the accuracy of bearing observation by direction -finder on a ship at sea is less than that Smith -Rose: Radio Direction -Finding 473

obtainable with a similar installation on land. The difference is, in fact, sufficient to make the optimum accuracy of the ship direction -finder comparable with that of the rotating beacon system. The direction -finding bearing is taken relative to the direction of the ship's head, and its accuracy depends upon the steadiness of the ship and also upon the accuracy with which the direction of the ship's head is given by the compass reading at any desired instant. In anything but a calm sea it is difficult to keep the head of a small ship steady to within one or two degrees, and in a rough sea the swing may amount to five degrees or more. Also there is probably a lag between the actual direction of the ship's head and the reading of the compass, whereas no such lag exists on the direction -finder. The bearing obtained from the rotating beacon is entirely free from these limitations and its accuracy is practically the same whether the ship is at sea or in dock.Also since the accuracy of observed bearing has been shown to be largely immune from conditions local to the receiver, no correction or compensation is necessary corresponding to the quadrantal error associated with the ship direction - finder. This error is likely to be serious in the case of many small ships, and there is the added possibility of its alteration with draught and nature of cargo. The limitation of range of accurate bearings due to night effect has been shown, both theoretically and experimentally, to affect both systems of directional wireless to the same degree. The accuracy of the rotating beacon has also been found to be sensibly the same for both continuous wave and interrupted continuous wave transmission.Considering the possibility of interference of the rotating beacon transmitters with other wireless services, a similar objection may be raised to the fixed beacon transmitter which has been found to be so necessary in conjunction with the ship direction -finder.In neither case, however, is any serious interference likely to be caused since definite and distinct wavelengths have now been allocated to both types of beacon. A small point just worthy of mention in concluding this section is that rotating beacons erected on the coast would be of some considerable service to aircraft navigators, particularly those making daily flights along the principal air routes. (d) Conclusions It will be gathered from the above discussion that the rotating beacon system is considered to have some advantages 474 Smith -Rose: Radio Direction -Finding over the ship's direction -finder in the application of directional wireless to marine navigation. These advantages are, however, probably not sufficiently well defined at present as to make the future policy of those concerned with these matters independent of certain other factors; and in any case it is evident that the economics of the situation must be considered as well as the technical features.The rotating beacon system would at first appear to be an economy in that the whole cost of the directional apparatus is removed to the authority controlling the shore stations; and it is to be presumed that the cost of installing and maintaining the rotating beacons would be met by the levy of a tax upon all ships using the beacons. This might be a great advantage to the shipowners who, judging from their tardiness in equipping their ships with direction -finders, do not appear to find this apparatus very cheap. To the cost of such direction- finders, moreover, is to be added that of the fixed beacons specially built for their use. This cost depends upon local conditions and whether or not the beacon can be installed at an existing lighthouse and maintained by the personnel of the lighthouse. The first rotating beacon for practical working to ships has yet to be built so that details of the cost are at present unknown. 7. Acknowledgments This communication is published by permission of the De- partment of Scientific and Industrial Research, London, Eng- land.The whole of the work described in this paper has been carried out for the Radio Research Board of that Depart- ment, and full acknowledgment must be made of the assistance rendered by the author's associates working for the Board in the various phases of the whole investigation. Bibliography 1(a). Fleming, J. A. The principles of electric wave telegraphy and telephony. Pp. 651-671, 1910. 1(b). Zenneck, J.Wireless telegraphy (English Trans.).Pp. 338- 370, 1915. Eccles, W. H. Wireless telegraphy and telephony. Pp. 117-120, 134-1361(c).,444-449,1918. 1 d). Walter, L. H. Directive wireless telegraphy. 1921. 1 e). Keen, R. Wireless direction finding and directional reception. (2nddition) 1927. l(f). Mesny, R. Usages des cadres et radiogoniometrie. 1925. 1(g). Long, S. H. Navigational wireless.1927. 2. Pickard, G. W. Determination of wireless wave -fronts. Elec. Rev.. (New York) pp. 494-495, Oct. 1908. Smith -Rose: Radio Direction -Finding 475

3. Fessenden, R. A. The Fessenden pelorus. Electrician 83, pp 719- 721; 1919. 4. Taylor, A. H. Variation in direction of propagation of long electromagnetic waves. Bureau of Standards. Scientific Ppr. No. 353; 1919. 5. Blondel A. On the goniometric functions applicable to directive aerials. Radio Rev. 1, p. 123; 1919. 6. Robinson, J.A method of direction finding of wireless waves. Radio Rev. 1, p. 271; 1920. 7. Mesny, R. The diffraction of the field by a cylinder and its effect on directive reception on board a ship. Radio Rev. 1, pp. 532-540, 591- 597; 1920. 8. Eckersley, T. L.Refraction of electric waves.Radio Rev. 1, pp. 421-428; 1920. 9. Kiebitz, F.New experiments with Scheller's directional transmitter. Jahrb. d. T. u. T. 15, pp. 299-310; 1920. 10. Hollingworth, J., and B. Hoyle. Local errors in radio direction - finding. Radio Rev. 1, pp. 644-649; 1920. 11. Round, H. J. Direction and position finding. Jour. I. E. E. 58, pp. 224-247; 1920. 12(a). Buchwald, E., and R. Hase.Directional reception experiments in aeroplanes. Jahrb. d. T. u. T. 15, pp. 101-113; 1920. 12(b). Buchwald, E. Experiments with Scheller's radio course setter on aeroplanes. Jahrb. d. T. u. T. 15, pp. 114-122; 1920. 12(c). Baldus, R., and E. Buchwald. Experiments on the wireless orientation of aeroplanes. Jahrb. T. u. T. 15, pp. 214-236; 1920. 12(d). Baldus, R., and R. Haim. Measurement of energy radiated by an aeroplane antenna. Jahrb. d. T. u. T. 15, pp. 354-391; 1920. 13. Burstyn, W. Wireless telegraphy in space. Jahrb. d. T. u. T. 16, pp. 322-337; 1920. 14. Adcock, F. Improvement in means for determining the direction of a distant source of electromagnetic radiation.British Patent No. 130490. 15. Kinsley, C., and A. Sobey. Radio direction changes and variation of audibility. PRoc. I. R. E. 8, pp. 299-323; August, 1920. 16(a). Smith -Rose, R. L.Variations of apparent bearings of radio transmitting stations.Part I.Radio Research Board.Special Report No. 2, 1924. 16(b). Smith -Rose, R. L.Variations of apparent bearings of radio transmitting stations.Part H. Radio Research Board.Special Report No. 3, 1928. 16(c). Smith-Rose' R. L.Variations of apparent bearings of radio transmitting stations. Part III. Radio Research Board. Special Report No. 4, 1926. 16(d). Smith -Rose, R. L. A study of radio direction -finding. Radio Research Board. Special Report No. 5, 1927. 17. Smith -Rose, R. L. On the electromagnetic screening of a triode oscillator.Proc. Phys. Soc. 34, pp. 127-138; 1922. 18: Barfield, R. H.Some experiments on the screening of radio receiving apparatus. Jour. I. E. E. 62, p. 257; 1924. 19. Smith -Rose, R. L., and R. H. Barfield. The effect of local conditions on a radio direction -finding installation. Jour. I. E. E. 61, pp. 179- 191; 1922. 20. Smith -Rose, R. L., and R. H. Barfield. The effect of undergound metalwork on radio direction -finders.Wireless World 11, pp. 165-171; 1922. 21. Smith -Rose, R. L., and R. H. Barfield. A discussion of the practical systems of direction -finding by reception.Radio Research Board. Special Report No. 1, 1923. 22. Marconi, G. Radio communication. PROC. I. R. E. 16, pp. 40- 69; January, 1928. 476 Smith -Bose: Radio Direction -Finding

23. Franklin, C. S.Short wave directional wireless telegraphy. Jour. I. E. E.60, pp.930-934; 1922. 24(a). White, R. H. The theory of the wireless beam.Electrician 94, pp. 392-393; 1925. 24(b). Dunmore, F. W.,and F. H. Engel.Directive radio transmission on a wavelength of 10 metres. Bulletin of Bureau of Standards Scientific Paper No. 469 19, pp. 1-16; 1923. 25(a). White, R. H. The wireless beam. Electrician 94, pp. 424-425; 1925. 25(b.) The wireless lighthouse. Wireless World 17, pp. 348-350; 1925. 26. Buchwald, E.Experiments with Scheller's radio course setter on aeroplanes. Jahrb. d. T. u. T. 15, pp. 114-122; 1920. 27(a). Kiebitz, F.New experiments with Scheller's directional transmitter. Jahrb. d. T. u. T. 15, pp. 299-310; 1920. 27(b). Kiebitz, F.Directive wireless telegraphy.Telegraphe u. Fernsprech-Technik. 9, pp. 46-50; 1920. 28. Kiebitz, F.The Telefunken compass.Electrician69, p.270; 1912. 29. Erskine -Murray, J., and J. Robinson.Directional transmission of electromagnetic waves for navigation purposes.Jour. I. E. E.60, pp. 352-356; 1922. 30. Gill, T. H., and N. F. S. Hecht. Rotating loop radio transmitters, and their application to direction -finding and navigation. Jour. I. E. E. 66, pp.241-255; 1928. 31. Engel, F. H., and F. W. Dunmore. A directive type of radio beacon and its application to navigation. Bulletin of Bureau of Standards Scientific Paper No. 48019, pp.281-295; 1924. 32. Dellinger, J. H.,and H. Pratt. Development of radio aids to air navigation.PROC. I.R. E.16, pp.890-920; July, 1928. 33. Pratt, H.Apparent night variations with crossed -coil radio beacons.PROC. I.R. E.16 pp.652-657; May, 1928. 34. Jolliffe, C. B., and E. M. Zandonim.Bibliography on aircraft radio.PROC. I.R. E.16, pp.985-999; July, 1928. 85. Kolater, F. A., and F. W. Dunmore. The radio direction -finder and its application to navigation. Bureau of Standards Scientific Paper No. 428, 17 pp. 529-567; 1922. 36(a). Dunmore, F. W. A uni-control high -frequency radio direction- finder. Bureau of Standards Scientific Paper No.525,21, pp.25-35; 1926. 36(b). Dunmore, F. W. A portable direction -finder for 90 to 7,700 kilocycles. Bureau of Standards Scientific Paper No.536,21, pp.409-430; 1926. 37. Smith -Rose, R. L. A sensitive long -wave radio direction -finder. Journal of Scientific Instruments 4, pp. 252-262. 38. Robinson, J. A method of direction -finding by wireless waves, and its applications to aerial and marine navigation. Radio Rev. 1, pp. 213- 219, 267-275; 1920, 39. Bainbridge -Bell, L.Marine direction -finding.Electrician 97, pp. 125-126; 1926. 40. Bellini, E., and A. Tosi.A directive system of wireless telegraphy. Elec. Eng.2, pp.771-775; 1907. Elec. Eng. 3, pp. 348-351; 1908. Proc. Phys. Soc.21, pp.305-326; 1908. Phil. Mag.16, pp.638-657; 1908. 41. Horton, C. E.Wireless direction -finding in steel ships.Jour. I. E. E.61, pp.1049-1056; 1923. 42. Slee, J. A. Development of the Bellini-Tosi system of direction- finding in the British mercantile marine. Jour. I. E. E.62, pp.543-550; 1924. 43. Mesny, R. Radiogoniometry on steel ships.Bull. Soc. Franc. Elec. 4, pp. 833-834; 1924. 44(a). Smith -Rose, R. L., and S. R. Chapman. An investigation of a rotating radio beacon. Radio Research Board.Special Report No. 6, 1928. Smith -Rose: Radio Direction -Finding 477

44(b). Smith -Rose, R. L., and S. R. Chapman. Some experiments on the application of the rotating beacon transmitter to marine navigation. Jour. I. E. E. 66, pp. 256-269; 1928. 45. Sommerfeld, A. The reciprocal theorem in wireless telegraphy. Zeit& f. Hochfrequenz. 26, pp. 93-98; 1925. 46. Smith -Rose, R. L. On the reversibility of radio direction -finding and local error at rotating loop beacons. Jour. I. E. E. 67,1929. 47(a). Appleton, E. V.The propagation of radio waves over the earth's surface. Nature 115, p. 382; 1925. 47(b). Appleton, E. V., and M. A. F. Barnett. A note on wireless signal strength measurements during the solar eclipse of 29th January, 1925. Proc. Camb. Phil. Soc. 22, p. 675; 1925. 48. Wright, G. M., and S. B. Smith. The heart -shaped polar diagram and its behaviour under night variations.Radio Rev. 2, pp. 394-403; 1921. 49. Eckersley, T. L. The effect of the Heaviside layer on the apparent direction of electromagnetic waves.Radio Rev. 2, pp. 60-65, 231-248; 1921. 50. Appleton, E. V., and M. A. F. Barnett. On some direct evidence for downward atmospheric reflection of electric rays.Proc. Roy. Soc. 109, pp. 821-641; 1925. 51(a). Hollingworth, J.The propagation of radio waves.Jour. I. E. E. 64, pp. 579-589; 1926. 51(b). Hollingworth, J. The polarization of radio waves. Proc. Roy. Soc. 119, pp. 444-464. 1928. 52. Smith -Rose, It. L., and R. H. Barfield. An investigation of wireless waves arriving from the upper atmosphere. Proc. Roy. Soc. 110, pp. 580-614; 1926. 53. Austin, L. W. Report of the chairman of the commission on radio wave propagation, U.R.S.I. PROC. I. R. E. 16, p. 356; March, 1928. 54. Merritt, E., C. C. Bidwell, and H. J. Reich. Changes observed in the direction of radio signals at the times of the eclipse of 24th January, 1925. Jour. Franklin Inst. 199, pp. 485-492; 1925. 55. Bidwell, C. C.Direction and intensity changes of radio waves. Jour. Franklin Inst. 201, pp. 107-112; 1926. 56. Reich, H. J. Variation of intensity and direction of radio signals. Jour. Franklin Inst. 203, pp. 537-548; 1927. 57. Pickard, G. W. The direction and intensity of waves from Euro- pean radio -telegraphic stations. PROC. I. R. E. 10, pp. 161-174; 1922. 58. Mesny, R. The variation and intensity of the electromagnetic field of a wave. L'Onde Electrique 1, pp. 501-517, 577-587; 1922. 59. Gherzi, E. Radiogoniometric observations at Shanghai. L'Onde Electrique 3, pp. 542-547; 1924. 60.Wireless observations during the eclipse of the sun, 29th June, 1927. Radio Research Board. Special Report No. 7,1928. 61. Ferric, G., R. Jouaust, R. Mesny, and A. Perot.Radiogonio- metric studies. Comptes Rend -us 172, pp. 54-58; 1921. 62. Mesny, R. The deviations of electromagnetic waves. Journal de Physique 4, pp. 129-143; 1923. 63(a). Stoye, K. Atmospheric conditions and electric waves. Jahr b d. T. u. T. 19, pp. 58-72; 1922. 63(b). Stoye, K.Wireless telegraphy and atmospheric conditions. Jahrb. d. T. u. T. 23, pp. 87-88; 1924. 64. Austin, L. W. A suggestion for experiments on apparent radio direction -finding variations.13 Hoc. I. R. E. 13, pp. 3-4; February, 1925. 65. Smith -Rose, R. L. Some radio direction -finding observations on ship and shore transmitting stations. Journal 62, pp. 701-711; 1924. 66. Rothe, E. Radiogoniometry and atmospheric influences. Comptes Rendus 172, pp. 1345-1347; 1921. 67. Smith -Rose, R. L. The effect of wave -damping in radio direction - finding. Jour. I. E. E. 63, pp. 923-927; 1925. 478 Smith -Bose: Radio Direction-Finding

68(a). Smith -Rose, R. L.On coastal errors in wireless finding. Nature 116,pp. 426-427; 1925. direction - 68(b). Barfield, R. H. Coastal 116, pp. 498-499; 1925. refraction of wirelesswaves. Nature 69. Baumler, M., and J. Zenneck. of electromagnetic Experiments on the propagation waves. Zeits. f. Hochfrequenz. 27,pp. 117-119; 1926. 70. Worlledge, J. P. G.Deviation of wirelesswaves at a coastal bo.undary. Nature 121,.p. 35;1928. 71. Slee, J. A.The progress of wireless mercantile marine. telegraphy in the British p. 819; 1924. Year Book of WirelessTelegraphy and Telephony. 72. Smith -Rose, R. L. A aerial arrangements for rotatingtheoretical discussion ofvarious possible pp. 270-274; 1928. beacon transmitters. Jour. I.E. E. 66, 73. Wright, G. M. Direction-finding. Year Book of Wireless raphy and Telephony.p. 953; 1920. Teleg- 74. Smith -Rose, R. L., and R. of night errors in radio H. Barfield. Thecause and elimination 1926. direction finding. Jour. I. E.E. 64, pp. 831-838; 75(a). Putnam, G. R. U. S.lighthouse service: radio fog Jour. Washington Acad. Science12, pp. 279-288; 1922. signals. 75(b). Putnam, G. R.Radio fog signals and the Dept. of Commerce, U. S.A., 1924. radio compass. 75(c). Slee, J. A. The problem 15, pi). 330-334; 1925. of beacon stations.Wireless World 75(d) Slee, J. A. Wireless beacons. 75(e). Slee, J. A. The Round Electrician 99, p. 251; 1927. pp. 600-601; 1927. Island radio beacon.Elec. Rev. 101, 75(f). Blondel, A. The Mi. 15, pp. 478-491; 1926.new French radio beacons. Ann. desPosies Proceedings of the Institute of Radio Engineers Volume 17, Number 3 March. 1959

SOME EXPERIMENTS IN SHORT DISTANCE SHORT-WAVE RADIO TRANSMISSION*

BY J. K. CLAPP (Communication Laboratory, Maaeachusetta Institute of Technology, Cambridge, Mass.)

Summary-Some experiments in short-wave radio transmission over a distance of 55 miles are described, the results of which are interpreled to indicate the presence of strong "sky" waves, with "ground" waves of negligible amplitude in comparison with the "sky" waves as received. Upon decreasing the transmitter wavelength, at a given time of day, a minimum wavelength was reached below which no communication could be obtained; this wavelength is termed the "cut-off" wavelength. The average value of the cut-off wavelength, for various times of day, is given for several different months. The minimum observed wavelength upon which communication was possible was 28 meters. A series of experiments in which an orientable half -wavelength antenna was employed served to indicate definitely an optimum position of the antenna for transmission over the 55 -mile distance. The indicated transmission path left the transmitter at an angle of approximately 65 degrees to the horizontal. In long distance communication the position of the antenna was found to have no appreciable effect.

PTHE work herein described was carried out over a period of two years, over a fixed distance of approximately 55 miles between Round Hill,South Dartmouth, Mass. and Auburndale, Mass. on wavelengths ranging from 25 to 80 meters. In view of the results obtained which are explainable ina rather straight -forward manner on the basis of refraction of the radio waves in the upper atmosphere, the presence of such a refracting medium will be assumed as an hypothesis.It was early found in the course of the experiments that at times no communication could be obtained on the wavelengths, and at the time of day, when under normal conditions very strong signals would be received. This phenomenon is believed to be due to conditions of the refracting layer such that it was impossible for refracted waves to return to the surface of the earth within the fixed distance of 55 miles. The noticeable absence of signals during these periods indicated that the "ground" signals, due towaves Dewey decimal classification: R113. Original manuscript received by the Institute, November 7, 1928. Contribution from the Round Hill Short-wave Radio Research, sponsored by Col. E. H. R. Green. 479 4S0 Clapp: Short Distance Short -Wave Radio Transmission following the surface of the earth, did not play any appreciable part in contributing to the usual received signal. A very complete theoretical study of the transmissions paths of short waves, through an electron atmosphere, has been made by Baker and Rice.' Their conclusions are here utilized in brief form, in the summary of the following paragraphs. For further interesting and valuable discussions of the problem the reader is referred to the books and papers below.2 Fig. 1 is a sketch showing the paths of the variousrays radiated by an ideal transmitter having a characteristic as sketched

/ ItI / --- --.11c - - ' I ...;-:."-:::"----- T - , .,.....-....., -.1 -.... 'I i/-4.- 1vt-4---- II`'`a stancr Ki.k: \\11,

CIRMCAL RAY TANGENT RAY SNIP DISTANCE

CRITICAL TRANSMISSION 6:1Z%` ANGLE (la) Fig. 1-Sketch Showing the Paths of the Various Rays Radiated byan Ideal Transmitter Having a Characteristic as Sketched in Fig. la, as Given by Baker and Rice. in Fig. la, as given by Baker and Rice. The tangentray, TD, has the greatest range of all the rays radiated close to the horizontal. The critical ray, TG, traverses the minimum distance along the surface of the earth.Rays leaving the transmitter near the vertical may return to the surface of the earth at great distances from the transmitter, as TJ, TK. Rays whichpene- trate to the region of maximum ionization are given curvatures W. G. Baker and C. W. Rice, "Refraction of Short Waves in the Upper Atmosphere," Jour., A.I.E.E., 45, p. 535-539, June, 1926; also reprints, of a more detailed nature, of paper presented to A.I.E.E. in February, 1926. 2A. H. Taylor, "Relations between Height of Kennelly -Heaviside Layer and High Frequency Radio Transmission Phenomena," PROC. I. R. E., 14, 521-540; August, 1926.A. H. Taylor and E. 0. Hulburt, "The Propagation of Radio Waves over the Earth," Phys. Rev., 27, February, 1926.Also QST, October, 1925.H. W. Nichols and J. C. Schelleng, "Propagation of Electric Waves over the Earth," Bell Sys. Tech. Jour., VI -2, 215-235, April, 1925. P. 0. Pederson, "The Propaga- tion of Radio Waves," published by "Danmarks Naturvidenskabehge Samfund," 1927, and sold by G. E. C. Gad, Vimmelakaftet 32, Copen- hagen, K. Denmark. (In English.) Clapp: short Distance Short -Wave Radio Transmission 481 departing only slightly from that of the layer itself; the rays may consequently pass around the earth before returning to the surface or may be bent out into space. Rays emitted at angles very near to the vertical pass through the layer and outinto space with relatively small changes in direction. In Fig. 1 an ideal transmitter is assumed at T, i.e., a system which radiates equally well in all directions above the surface of the earth- as sketched in Fig.1A. Considering now the range of rays emitted at various angles above the horizontal, we find that the tangent ray TD, if unaffected by surface conditions, will return to the surface of the earth at a distance, in round numbers, of 3,000 miles from the transmitter T.This distance is only slightly dependent upon the wavelength and the conditions of the refracting layer, because of the relatively small portion of the path which is traversed within the refracting layer.For increasing angles of transmission, the range is decreased, very rapidly at first and then more and more slowly until the minimum range is reached (TG), at which pointthe change of range as a function of transmission angle is zero. The path for the minimum range is very critically influenced bythe conditions of the refracting layer (height, thickness, value of maximum ionization density and the distribution of ionization), because of the greater penetration and the greater curvature given the ray, and consequently the path varies rapidly with the wavelength.If higher angles of transmission are considered, the range increases as the angle is increased, slowly at firstand then more and more rapidly. At the higher angles, the rays might travel around the earth before being given a sufficient curvature to return them to the earth's surface.For angles near the vertical, the rays are bent but little, and pass through the layer and out into space. The particular point of interest in the summary above is that, with fixed layer conditions, there exists a minimum range for the refracted rays, associated with a definite transmission angle. This minimum range is many times confused with the "skip - distance," i.e., that range over which no signals at all are received. When quite short waves are used, so that the minimum range is of the order of several hundred miles, this confusion is not serious, because of the usually limited range of the ground wave. For short distance operation, however, the ground wave intensity becomes very important in determining the actual skip -distance. The observed skip -distance, in these cases, depends not only on 482 Clapp: Short Distance Short -Wave Radio Transmission the minimum range of the refractedwaves (which defines the outer skip -distance boundary) but also on the maximumrange of the ground waves (which defines the inner skip -distance boundary). Any factors tending to decrease the first,or to increase the second, will result in a decreased skip -distance.If the ground waves reach the minimum refractedwave range with an appreciable intensity, then no skip -distance is encountered, as, for example, in short-wave transmission from aircraft.If the surface waves are highly attenuated near the transmitter, then the only recourse for short distance short-wave communication is to utilize relatively high angle rays. For certain layer conditions, short distance communication will not be possible unless the power of the transmitter is increased to a point where the attenuation of the groundwaves is overcome by sheer force to provide a signal of practical intensity. In these experiments the power output of the transmitterwas purposely restricted to 50 watts, or less, in order that good reception would be attained only when favorable transmission conditions existed, making the delineation between favorable and unfavorable conditions as marked as possible.In practice there is little difficulty in identifying the "ground"wave and "sky" wave signals as received.The former entirely lack the general characteristics of short-wave transmission, i.e., violent and rapid changes in amplitude (even when the transmitter is carefully controlled as regards frequency and amplitude of the antenna current) ; the ground wave signals are remarkably steady, reminding one of very long wave signals. For long range communication, 5,000 milesor more, it is seen from the considerations above that two classes of transmission path are possible. The first employs therays radiated close to the horizontal at the transmitter, which are effective to distances of approximately 3,000 miles. To coverranges greater than this it is necessary to invoke the aid of multiple reflections from the earth's surface. The idea is theoretically tenable, but has no great appeal from a practical viewpoint owing to the high absorption encountered by the waves at surface grazing regions, and the scattering effect of surface irregularities.The second class of path is provided by any rays which cover a range greater than 3,000 miles, invoking no multiple reflections. These paths are the high angle paths, indicated by the dashedcurves of Fig. 1. The loss of energy due to dispersion of these pathsover the Clapp: Short Distance Short -Wave radio Transmission 483 earth's surface may in a large measure be overcome by the low attenuation, resulting in signal amplitudes of practical interest. The experiments of Meissner' with a "searchlight" beam transmitter definitely indicate the usefulness of comparatively highangleradiationineffectingcommunication between Germany and Argentina on a wavelength of 11 meters. Two transmission angles for maximum received signal strengthwere found within the range of the observations, one at approximately 80 degrees and the other at approximately 37 degrees with the horizontal. It is regrettable that the physical limitations of the reflectors did not permit of swinging the beam to the horizontal, instead of a minimum of 30 degrees. Transmission at high angles

se AO

CRITICAL RAYS ceern SCIltrsez.

Fig. 2-Paths of the Critical Rays (TG of Fig. 1) for Various Wavelengths, Based on the Work of Baker and Rice. over great distances is further substantiated by the general experiences of amateurs, using low power, and, in the majority of cases, essentially horizontal antennas at such heights above earth that a strong component of radiation is sent out at high angles. Returning to the specific problem of short -distance communication, where the distance to be covered was sometimes less and sometimes more than the minimum refracted ray range, observations were made to determine the minimum wavelength which would provide reliable reception at various times of day and in various seasons. With given conditions of the refracting layer, the minimum range of transmission will increase as the wavelength is decreased, somewhat as sketched in Fig. 2. For a given distance between transmitter and receiver, T-R, a wavelength of 40 meters would produce strong signals, while a wavelength of 35 meters would fail entirely. For some wavelength, A. Meissner, "Directive Radiation with Horizontal Antennas," PROC. I. R. E., 15, 930-934; November, 1927. 484 Clapp: Short Distance Short -Wave Radio Transmission here about 38 meters, the path of the minimum range ray would strike the earth in the immediate vicinity of the receiver. Any slight irregularities of the layer conditions would then cause the foot of the path to sweep back and forth across the receiving point, producing violent "fading" or "fluttering."Such was found to be the case in these tests, the signals swinging from relatively large amplitudes to zero. Under more stable conditions of the layer, the fluttering was not pronounced, and, under such conditions, the wavelength at which communication just failed or was "cut off," could be determined within about one per cent. Fig. 2 shows the paths of the critical rays (TG of Fig. 1) for various wavelengths, based on the work of Baker and Rice. For each wavelength the distance from the transmitter T to the point at which the ray returns to the earth's surface is approximately the "skip -distance."For the conditions of the diagram, wave -

TRANSMITTER RECEIVER

p...._1-MISSMP1-37 SUPPLY tVARIABLEF1 I 25-8D 64. Fig. 3-Arrangement of Duplex Transmitter -Receiver Equipment Used in Making the Observations Described.

lengths longer than about 38 meters will be received at R, but wavelengths shorter than this value will not be received. The equipment was arranged as in Fig. 3.Channel No. 1 was arranged as a communication channel, or "order wire," to be operative on wavelengths which had been found to give reliable communication at the time of day at which the experiments were being conducted.Channel No. 2 was arranged with a special transmitter designed for rapid manual change of wavelength. The transmitters were keyed simultaneously on the same power supply. The receivers were of similar design, consisting of an autodyne detector followed by two stages of audio -frequency amplification. Each receiver connected to one telephone receiver of a "split" headset.'If equal signal intensities were obtained on both wavelengths, and if the operator adjusted the autodynes to produce the same beat tone, the signal heard in the headset was remarkably similar to that obtained with a single receiver, except that the signal appeared to change from one ear to the J. K. Clapp, "Multiplex Short Wave Reception," QST, March, 1926. Clapp: Short Distance Short-Yave Radio Transmission 485 other, as the fading periods on the two wavelengths were not the same. No definite relationship between the fading periods on any two wavelengths was observed, even when the difference in wavelength was small.'Ordinarily the fading period, on the wavelengths covered in this investigation, is so short that the effect is that of an irregular amplitude modulation of the signal

TIMM - Fig. 4-Illustrating Average Variation in Received Signal Strength, for Short Distance Communication, as a Function of the Time of Day. as received, made evident to the ear by a roughening of thebeat tone without much variation of the average beat tone intensity. By changing the wavelength of transmitter No. 2 in successive steps, a general idea of the "cut-off" wavelength at a given time of day and under given conditions of the refracting layer could readily be obtained. The received signal was found to vary with

A. tames-. 30 35 40 46 SO Fig. 5-Illustrating the Variation in Received Signal Strength, for Short Distance Communication, at a Given Time of Day, as a Function of the Wavelength of the Transmitter. the wavelength as sketched in Fig. 5. When the signal had been reduced to a point where it could just be identified, though far too weak for communication, communication was considered as being "cut-off."Reception varied with the time of day, on a R. Brown, De L. K. Martin, and R. K. Potter, "Some Studies in Radio Broadcast Transmission," PROC. I. R. E., 14, 57-131; February, 1926. 486 Clapp: Short Distance Short -Wave Radio Transmission

fixed wavelength, as sketched in Fig. 4.The values shown in Figs. 4 and 5 are for summer conditions. A decided drop in signal strength accompanied any marked increase in wavelength above the cut-off value. It was soon found, in making observations, that the value of the cut-off wavelength was subject to large and rapid fluctuations.' As interest was here centered on average values, no attempt was made to develop means for making more rapid determinations. The oscillographic method described by Heising could well be adapted to this study, as well as the method of "wobbling" the transmitter frequency.' 111111111111E 11111111111111

30 AVERAGE C.URVE APRIL 1926

11111"10 MON 2 4 6 6 MER. TIE PIA Fig. 6

The results obtained, over the fixed distance of 55 miles, are set forth in Figs. 6 to 11 inclusive. The early summer and mid- summer records show a rather flat curve, while the late fall records show a more or less regular steep curve.The cut-off wavelengths are plotted against the time of observation, all times being given in Eastern Standard Time. Signals on wavelengths lying just below the curve were not received, while, in general, good reception was obtained on wavelengths lying just above the curve; that is, on wavelengths just above the cut-off wavelength. Observations were made on three dates in April, 1926, between the hours of 9 A.M. and 1 A.M. of the succeeding day. The curve of Fig. 6 shows the minimum wavelength on which communication was possible in this period to have been 32 meters, occurring at about 5 P.M. ° R. A. Heising, "Experiments and Observations Concerning the Ion- ized Regions of the Atmosphere," PROC. I. R. E., 16,75-99; January, 1928. 7 R. A. Heising, J. C. Schelleng, and G. C. Southworth, "Some Measurements of Short Wave Transmission," PROC. I. R. E., 14, 613; October, 1926. Clapp: Short Distance Short -Wave Radio Transmission 487

Fig. 7 indicates the results of tests during May, 1926, in which efforts were made to obtain data throughout the twenty-four hours. From 9 A.M. to 1 A.M. definite results were obtained, but from 1 A.M. until 9 A.M. particularly between 5 A.M. and 9A.M.,

40 Ild 11111,1111

MAY 18211

i2 2 4 6 6 10 NOON 2 4 6 8 10 12 Am 75. MEP. TIME PM Fig. 7 the results were quite frequently indefinite.In every case the value of the cut-off wavelength rose abruptly just before sunrise, reaching a maximum shortly after sunrise and then gradually diminishing throughout the morning. The minimum wavelength on which satisfactory communication could be obtained during this period was 31 meters, attained at approximately 5 P.M.

-4....3

9. 30.

... JULY11126

a 2 4 6 6 10 NOON 2 4 6 6 10 12 AM 73" MER. TIME PM Fig. 8 Fig. 8 shows the results of a three-day test in July, 1926. During this period the signals obtained after midnight did not show any evidence of being due to refracted waves, but appeared to be due solely to the ground wave, as in this interval the signals were characterized by extreme constancy of amplitude. In the interval from 6 A.M. to midnight the results were quite definite, as shown by the curve.The minimum wavelength for communication in this interval was about 32 meters. 488 Clapp: Short Distance Short -Wave Badio Transmission

At the end of July, 1926, a four -day test was made, after improvements in the methods of handling the tests had made possible more rapid determinations. The four curves of Fig. 9 show the great fluctuations in the value of cut-off wavelength which were encountered in relatively short intervals of time. The

1 ze" I I 50 4i.

40 N.A.CI#rdr O30

3 JULY 1926

4 6 .0 .2 .2 2 4 6 13 10 NOON 2 8 75" MEP. 16.1E Fig. 9 average cut-off wavelength during these tests was higher than that obtained earlier in the month, but at times wavelengths very near to 30 meters were observed. The erratic nature of these curves is typical of the observations by the more rapid method, but it is surprising what a regular curve is obtained on averaging a large number of such observations.

....,,,....

4 .

MaS ) NERAGE CURVE JULY -AUGUST

I 1"4l .2 2 2 4 6 8 10 NOON 2 4 _6.. PI .0 AM 75 MEFI. TIME Fig. 10 All of the data for the tests of July and August, 1926, were averaged and the curve of Fig. 10 shows the result. The minimum wavelength of the average curve is 37.5 meters, where under temporary and more favorable conditions wavelengths as short Clapp: Short Distance Short -Wave Radio Transmission 489

as 31 meters were observed. The results of this averaging are so striking as to suggest strongly that the average conditions of the refracting layer undergo a regular variation. From time to time, "bumps" or "hollows," or regions of greater or less density of ionization, drift across the path of transmission, causing the transient irregularities which were observed. Fig. 11 shows the average of the observations for October, 1926, the minimum wavelength observed being 29 meters at approximately 1P.M. The curve as a whole has steepened materially, indicating that the number of hours of the day during which satisfactory communication could be obtained was de- cidedly less than in the preceding months. It would be expected

60,

I50

40 1

30 AVERAGE CURVE OCTOSER 192$

I I I 2 4 6 6 10 MOON 2 4 6 6 10 12 AM /5" MER. TIME PM Fig. 11

that the minimum value of cut-off wavelength would increaseas winter came on, owing to the reduced density of ionization,or to an increased average height of the refracting layer. The reduction in cut-off wavelength here encountered may be associated with the heavily overcast weather which was encountered during the period of the observations. In this connection it was repeatedly observed that when large, well-defined clouds drifted across the path of transmission the received signals underwent a definite change in intensity. When the clouds were sufficiently well separated so that distinct observations could be made, it was found that the signal intensity rose when the cloud was in the line of transmission. On days which were heavily overcast, with or without fog at Round Hill (the station there being located within twenty yards of the shore of Buzzard's Bay), the average signal intensity was materially greater than on clear days.When rain was falling over the 490 Clapp: Short Distance Short -Wave Radio Transmission territory including both transmitter and receiver, the signals fell below the clear weather average.' At times when the cut-off wavelength changed very rapidly, as between 7 and 9:30 P.M. during October, 1926, it was impossible to change wavelength with sufficient rapidity by manual methods. Resort was then had to the use of two transmitters at each point, the wavelengths being "staggered." For example: Round Hill Auburndale No. 1 35 meters No. 1 37 meters No. 2 39 meters No. 2 41 meters As the cut-off wavelength rose toward 35 meters, the signals from Round Hill would begin to flutter and then die out. Auburn- dale then communicated with Round Hill on 37 meters, until the cut-off wavelength rose to that value. Round Hill, on the extinc- tion of their 35 -meter signals, would transmit on 39 meters, thereby maintaining communication. If time permitted, Round Hill would increase the wavelength of transmitter No. 1 to some value higher than 41 meters, thus extending the range of the observations. By this means it was possible to obtain the times at which each of the four, or more, wavelengths were cut off. In many instances it was found that the cut-off wavelength rose 10 to 20 meters above its initial value in a period of ten or fifteen minutes.In the early evening the cut-off wavelength often reached the vicinity of 55 meters (above which no sharply defined cut-off effects were observed) from a value of 38 to 42 meters within fifteen minutes.Communication then usually failed entirely, except for faint signals on wavelengths of 80 meters or more, until the following forenoon. When communication failed abruptly in this manner, no instance was observed in which communication could be re-established on the same wavelength during the same evening.It appeared as though the characteristics of the refracting layer altered abruptly and did not return to the average normal until the changes of the ensuing day had taken place. The above discussion of short distance transmission, including the interpretation on the basis on high angle refracted waves, is based on the theoretical premise of the work of Baker and Rice. In order to obtain an experimental verification of this viewpoint, 8 "The Effect of Water Vapor in the Atmosphere on the Propagation of Electro-magnetic Waves," Jahrb. d. Draht. Tel., 12, 184; August, 1927. Also 19, 58-72; January, 1922. Clapp: Short Distance Short -Wave Radio Transmission 491 a series of tests was conducted with an orientable transmitting antenna system.9 The antenna proper consisted of fifty feet of copper tubing mounted on a light wooden lattice frame, carrying the transmitter at the midpoint. The center of the frame was supported on a universal joint at the top of a fifty -foot telegraph pole, set "in the clear" on the beach at Round Hill. Thus it was possible to place the antenna in any desired direction in space, the center of the antenna system being fifty feet above the surface of the earth. The first series of tests was made to determine the best position of the antenna for transmission from Round Hill to Auburndale, using a wavelength of 42 meters.Definite results were very difficult to obtain because of fast fading; by graphically recording the received signals for a period of three minutes for each chosen position of the antenna, taking the average value of the record by planimeter, and taking a large number of observations, some indication of the average received signal intensity was obtained. An optimum position of the antenna was clearly indicated. The antenna was contained in a vertical plane passing through the transmitter and receiver, with the end of the antenna nearest to the receiver depressed some 25 deg. below the horizontal, as indicated in Fig. 12. Application of antenna theory to this configuration indicates that the antenna radiates with greatest intensity at an angle of about 65 deg. above the horizontal, in the direction of the receiver, and that the image of the antenna does not contribute materially to this radiation.The optimum position of the antenna indicates that the effective transmission path left the transmitter at an angle of 65 deg. to the horizontal, bearing out the idea of high angle transmission. Consideration of the antenna and its image would lead one to expect fairly good transmission if the antenna were horizontal and at right angles to the line between stations; this was found to be true. .If the antenna were placed in the line of the transmission path, then the only received signal would be due to the image, which would of course depend largely upon the character of the earth. With highly comducting earth, good signals should be obtained, while with highly absorbing earth the signals would be weak.In a few observations moderately good signals were obtained; in a very large number the signals were quite weak and G. W. Pickard, "The Polarization of Radio Waves," PROC. I. R. E., 14, 205-212; April, 1926, and references given therein. 492 Clapp: Short Distance Short -Wave Radio Transmission in three instances no signals at all could be obtained. The experimental evidence is far from complete, but it indicates a condition of very imperfect reflection from the earth in the vicinity of the transmitter. The results of tests in which the angle of inclination of the antenna was varied, while the antenna remained in the plane containing the transmitter and receiver, are given in Fig. 12. The theoretical radiation characteristics of antennas in "principal" positions, i.e., horizontal or vertical, over perfectly

Iti

- 447 ; VALUES IN % or AVG. MAX. ...-8[amr ammo] JUL"' 1926. Fig. 12 conducting earth, have been considered by various investigators.w The problem of the antenna over an imperfect earth has been attacked also." The problem of an inclined antenna offers many difficulties, even if perfect earth conditions are assumed. In this 10 G. W. Pierce, "Electric Waves and Oscillations."S. Ballantine, "On the Radiation Resistance of a Simple Vertical Antenna at Wave- lengths below the Fundamental," PROC. I. R. E. 12, 823; December, 1924. S. Ballantine, "On the Optimum Transmitting Wavelength for a Vertical Antenna over Perfect Earth," PROC. I. R. E., 12, 833; December, 1924. S. A. Levin and C. J. Young, "Field Distribution and Radiation Resistance of a Straight Vertical Unloaded Antenna Operating at One of its Har- monics," PROC. I. R. E. 14, 675; October, 1926. 11 L. Bouthillon, "Influence de la nature du sol sur l'emission et la reception radio-electriques," L'Onde Electrique, 6, No. 71; November, 1927. Clapp: Short Distcinee Short -Wave Radio Transmission 493 case the earth surface was composed of fine sand, which was highly piezo-electric," showing strong resonances at frequencies from as high as 15,000 kc to less than 3,000 kc, that is, over a range of frequency considerably greater than that employed in the experimental work. One is then confronted with the vexatious though interesting problem of trying to account for the effect of the dancing sands under the atenna in attempting to calculate the radiation characteristics. When using this antenna system in long range communication, it was found that the position of the antenna had no marked effect, though no careful measurements could be carried out. Apparently the random polarization of the waves at the receiver, due to the characteristics of the medium of propagation, completely wiped out all effects due to the position of the antenna itself." These longer range tests were carried out mainly with amateurs in England and Belgium. In conclusion the author wishes to express his appreciation of the able assistance and active cooperation of Mr. Walter D. Siddall and Mr. Gordon G. Macintosh in carrying out the experiments. 12 I am indebted to Dr. G. W. Pierce for the suggestion that the piezo- electric character of the sand might be an important influence. 13 Baker and Rice, loc. cit. Proceedings of the Institute of Radio Engineers Volume 17, Number 3 March, 192.9

WIRELESS TELEGRAPHY AND MAGNETIC STORMS*

BY H. B. MARIS AND E. 0. HULBURT (Naval Research Laboratory, Bellevue, Anacostia, D. C.) Summary-A recent theory of auroras and magnetic storms attributes these phenomena to the action of a flash of ultraviolet light from the sun. The flash causes an unusual ionization in the Kennelly -Heaviside layer.There- fore, it is only daylight wireless circuits which are, or may be, disturbed at the commencement of the magnetic storm, the night circuits remaining normal until dawn when they may be disturbed; the disturbance in the daytime circuits may persist after night -fall.This very simple theory is found to be borne out in a detailed discussion of the data of the short-wave (15 to 40 meters) circuits of the United States Navy during the magnetic storms of May 28, July 7, October 18, and October 24, 1928. /T is an accepted belief, based on a long accumulation of evidence, that terrestrial magnetic storms are caused by unusual bursts of radiation of some sort from a disturbed region of the sun.The assumption that the solar disturbance is a flash of ultraviolet light has been shown in a recent paper' to explain practically all the complicated magnetic effects observed during a magnetic storm.In a simple case the flash was assumed to blaze out at full intensity for some minutes and then to die away in a few hours or days; it might, of course, flare out again irregularly or intermittently.The flash was such as would come from a hot spot 1/10,000 of the solar disk in size and at a temperature of 30,000 degrees K; the radiation from the hot spot would be largely in the far ultraviolet region of wavelengths which ionize the atmospheric gases.Calcula- tions showed that this radiation, barely perceptible at sea level because it is almost entirely absorbed in the upper reaches of the atmosphere, would heat up the atmosphere at levels above 50 km causing it to expand, and would increase the ionization in the high atmosphere very greatly.The effect of the flash was therefore to increase the ion and electron densities in the Kennelly -Heaviside layer and to raise the entire layer upward about 100 km.This agreed well enough with the measure- Dewey decimal classification: R113.5. Original manuscript received by the Institute, December 28, 1928. Published by the permission of the Navy Department, Washington, D.C. 1 H. B. Maris and E. 0. Hulburt, "A Theory of Auroras and Magnetic Storms," Phys. Rev., March, 1929. 494 Maris and Hulburt: Wireless Telegraphy and Magnetic Storms 495 ments of Dahl and Gebhardt2 who found an increase in the apparent height reached by 70-m waves in the daytime of about 70 km during the magnetic storm of August 19, 1927, and with the recent measurements of Hafstad and Tuve,' who observed an increase in the apparent daytime height of the layer of 100 km during the magnetic storms of October 7 and of October 18, 1928. The atmosphere on the daylight side of the earth is influenced directly by the flash, that on the night side is not affected directly but indirectly, as the day side is turned by the rotation of the earth and carries with it into the shadow the influence it received while in the sunshine.Therefore, only daytime wireless communication may be disturbed during the commencement of the magnetic storm, the night communication remaining normal.If the flash continues, with perhaps lessening intensity, the night communication will, or may, experience disturbance at dawn. If the communication channel extends from night into day, or vice versa, the effects may be complicated. Since the flash heats the atmosphere and thereby raises the Kennelly -Heaviside layer in the sunlit areas, long - wave daylight transmission may be improved, or probably will not be weakened, during the magnetic storm, because the waves are propagated with fewer earth and layer reflections than normally. At night, however, due to the diffusion of the unusual amount of ionization into lower strata where the gas molecules are more dense, the absorption of the long waves is increased and the transmission falls below the normal night- time values. These conclusions are in general accord with the facts brought out by Austen,* Espenschied, Anderson and Bailey,' Pickard,' Anderson," and others for waves from 5000 to 20,000 m. In general, long waves are not so sensitive as the short waves to movements and changes in the upper atmosphere, because of their length and their relatively less penetration into the ionized regions.Therefore, the very superficial description just given of the effects of the storm in the high atmosphere, 2 O. Dahl and L. A. Gebhardt, PROC. I. R. E., 16, 290; March, 1928. 3 Hafstad and Tuve, Terr. Mag. and Atmos. Elec., 34, March, 1929. ' L. W. Austin, PROC. I. R. E., 15, 825; October, 1927. 6 L. Espenschied, C. N. Anderson, and A. Bailey, PRoc. I. R. E., 14, 7; January, 1926. 6 G. W. Pickard, PROC. I. R. E., 15, 749; September, 1927. 7 C. N. Anderson, PROC. I. R. E., 16, 297; March, 1928. 496Maris and Hulbert: Wireless Telegraphy and MagneticStorms

which has perhaps sufficed fora rough explanation of the behavior of the long waves during the storm, is hardlysufficient for the case of the short waves. Amore careful examination is necessary, especially as dataare now available in considerable detail of the short-wave storm characteristics.The conditions in the atmosphere arising from the solarultraviolet flash are in the main complicated.If, as a result of the heavy ionization caused by the storm, the lower surface ofthe ionized layer is sharply defined, the shortwaves may be reflected without great penetration and their transmission willbe good; but if the lower surface is not well marked,or if it is agitated, etc., the transmission may be greatly disturbed.And which of the conditions obtains inany particular case depends upon the rates of diffusion of the ions and electrons,the winds in the high atmosphere, the character of the solar flash,e.g., whether it is short and intense, or mild and long,or irregular, etc., etc. Actually the various possibilitiesare met with, as brought out in the detailed discussion of four magneticstorms given in the following paragraphs, there beingcases in which theshort- wave full daylight transmission was greatly disturbedat the commencement of the storm, and othercases in which it was disturbed to a less extent.But, it must be emphasized again, no case was encountered in which the full night-timecommun- ication was troubled perceptibly at thecommencement of the storm.Since the atmosphere is heated and expandedat the beginning of the storm in those regions wherethe sun is more or less overhead, as from ten to four o'clock local solar time, winds in the high atmosphere will blow fromthe heated regions and produce turbulence and swirls of theions and electrons on the borders of those regions.Short wireless waves passing into this turbulence will be scattered,or absorbed, or poorly reflected, as the case may be, andpoor transmission will result. When night falls over the daylightareas, the turbulence may slowly diminish and the transmissionmay improve in the early morning hours, and then with the dawnmay continue to improve or may fall off again, dependingupon the activity of the flash.Thus, in part, the solar disturbance,which causes the magnetic storm, accentuates and widens theusual sunset and sunrise vagaries in the wirelesstransmission. In Figs. 1, 2, 3, and 4are given data for the four magnetic storms of May 28, July 7, October 18,and October 24, 1928. Maris and HuMart: Wireless Telegraphy and Magnetic Storms 497

respectively. The wireless information was obtained from the log books of the Radio Division of this Laboratory, and the magnetic data from the Coast and Geodetic Survey Observa- tory at Cheltenham, Maryland, U. S. A. The abscissas of the figures are the Eastern Standard times of the various days. The ordinates give the wireless communication as expressed qualitatively by the words "good", "poor", and "bad". "Good" means the normal signal intensity obtained with the usual day and night routine wavelength bands which were 14 to 25 and 25 to 40 meters, respectively; "poor" means signals below normal due to fading and low intensity, and "bad" means signals very much below normal, or absent altogether.Curves 1 of the figures refer to the short-wave communication from Wash- ington, D. C., U.S.A., to Europe, along the Atlantic sea -board and to South America.Curves 2 refer to the Washington -San Francisco short-wave circuit. Curves 3 of the figures give the variations in H, the horizontal component of the magnetic field of the earth; these curves are free -hand sketches from the magnetograph records of the Cheltenham Observatory. On a quiet day the H curves are smooth with diurnal undulations of a regular type. A ragged H curve with a sharp initial rise, a fairly rapid decrease to negative values and a slow period of recovery, is the characteristic type of storm curve. These characteristics are world- wide, being the same, apart from local variations peculiar to the station, for all the magnetic stations of the earth; the storm commences simultaneously, within a minute at all the stations. According to the theory of auroras and magnetic storms', the initial rise and fall to zero of the H curve is due to a pulse of current flowing from west to east around the earth in the conducting region of the atmosphere. The pulse of current is caused by the increased ionization in the Kennelly -Heaviside layer when the ultraviolet flash flares up to full intensity. The decrease of H to negative values is attributed to the induced currents in the earth, and the slow recovery of H to the dying away of the currents as well as to the fading out of the flash.Irregu- larities which occur during the recovery period probably indicate recurrences of the flash, although the induced current reactions may also play a part. The four storms illustrate four quite different types of conditions; the effects on wireless communication, however, were 498Maris and Hulburt: Wireless Telegraphyand Magnetio Storms

in each case in general accord with thetheoretical views which have been outlined. On May 27the magnetic disturbance began gently at about 10 A.M., as shown in Fig.1, increased to a moderate intensity on May 28 and diedaway the next day.

May27 May 28 Fig. 1

IV%

Fig. 2

Fig. 3

H 3 good poor bad Oct 21.1 240c12512 .4 Fig. 4 In the above figures the abscissasare Eastern Standard Time. Curve 1 gives the short-wave communicationon the Washington circuits to Europe and South America and along theAtlantic seaboard; curve 2 refers to the Washington -San Francisco circuit."Good" means the normal signal strength. Curve 3 gives H, the variationsin the horizontal component of the earth's magnetic field, as sketched fromthe magnetograph records of the Cheltenham Observatory, Maryland,U.S.A. Wireless showed no evidences of troubleon the 27th, but at about 6 A.M. of the 28th the wireless channelsboth to the east and the west became inoperative. Duringthe day of May 28 wireless was poor, and then began torecover in the night. The Maris and Hulburt: Wireless Telegraphy and Magnetic Storms 499 recovery on the Atlantic circuits continued through the dawn of May 29 and was complete by noon. The San Francisco communication, however, became poor again at sunrise of May 29 and did not return to normal until that night.Just why the San Francisco circuit experienced more trouble on May 29 than the Atlantic circuits is not clear. Perhaps local winds in high atmosphere, irregularities in the ionization, or many other factors neglected in the present simple theory were the cause. It should also be mentioned that the wireless data indicated in a number of cases a non -reciprocity of signalling, that is, signals were received at one station from a second, but the second station did not receive the signals from the first. How- ever, the data were not sufficiently complete, nor is the theory sufficiently detailed, to permit the discussion of these complex- ities.They will require further investigation. On July 7 at 4:30 P.M. a short severe magnetic storm set in and continued until about noon on July 8, as shown in Fig. 2. Communication from Washington both east and west dropped sharply to low values at about 7 P.M., and recovered slowly during the night and the following day.The Pacific circuits Honolulu to Cavite, which lay nearly directly under the influence of the storm flash, apparently were not disturbed on July 7 (actually July 8 at Cavite, which is on the other side of the international date line), but at 8 A.M., Honolulu local solar time, of July 8 short-wave communication faded to low values for ten hours. The storm of October 18, Fig. 3, began at 2:27 A.M., with a fairly pronounced magnetic disturbance; the Washington circuits to the east and west showed no indications until the dawn.During the daylight hours of the 18th the Atlantic communication was poor or zero; the San Francisco circuit, although somewhat troubled, remained in operation. A storm of moderate intensity, Fig. 4, began slowly in the morning hours of October 24 with a small increase at about 2 P.M.The Washington -San Francisco circuit had periods of low signal strength from 9 A.M. to 12, and in the afternoon communication became poor.The Atlantic circuits were little disturbed.Meanwhile the Pacific circuits, which were in the dark when the storm began on the 24th, experienced low signal strength and fading during the daylight hours of this period. In addition to the facts which have been mentioned, the 500Maris and Hulburt: Wireless Telegraphy and Magnetic Storms wireless log books indicated clearly that fading and poor signal strength were the general rule on days of magnetic character 1 or greater; 1 means moderate, 2 strong magnetic disturbance, etc.In 1927, a year of maximum disturbance, there were 83 days of character 1 or greater; wireless therefore might have encountered difficulties due to solar activity on more than one fifth of the days of the year. Three or four years from now the solar activity will be at the minimum of its eleven -year cycle and one may expect a less number, perhaps only 10 or 20, of magneti- cally disturbed days during the year. The magnetic observations may properly be regarded as the weather stations for the wireless traffic and the magnetic storm charts as the wireless weather maps. In conclusion it is a pleasure to express our thanks to the Coast and Geodetic survey for their courtesy in giving us promptly and fully the magnetic data of their observations. Proceedings of the Institute of Radio Engineers Volume 17, Number 8 March, 1929

RECENT DEVELOPMENTS IN SUPERHETERODYNE RECEIVERS*

BY G. L. BEER& AND W. L. CARLSON2 Westinghouse Electric and Manufacturing Company, East Pittsburgh, Pa.,' General Electric Company, Schenectady, N. Y.) Summary-Major electrical elements of a modern superheterodyne receiver -tuned radio -frequency amplifier, intermediate frequency amplifier detector and audio frequency amplifier-are briefly discussed in light of recent developments.A practical automatic volume control is described. Curves illustrating the major performance characteristics of the receiver are shown. - INTRODUCTION IN the past four years broadcasting conditions have changed so greatly that receivers which were giving satisfactory service at the beginning of that period are now obsolete. The steady increase in the number of broadcasting stations and the advent of the super -power stations have imposed exacting selectivity requirements on modern receiving sets.The congested condition of our cities with their numerous apartment buildings and the resultant lack of antenna facilities have likewise created the demand for receivers having sufficient sensitivity to permit the use of either a small indoor or outdoor antenna. The marked improvement in the quality of radio programs and their transmission and the development of the moving coil type of loudspeaker have made the fidelity of present day receivers a consideration of major importance. The superheterodyne receiver is particularly adapted to meet these modern broadcasting conditions. It is the purpose of this paper to discuss some recent developments in this type of receiver.

GENERAL The conventional tuned radio -frequency receiver has two main electrical elements, namely, the tuned radio -frequency amplifier and the detector and audio -frequency amplifier. The modern superheterodyne receiver has both these elements and, in addition, an intermediate frequency amplifier with associated *Dewey decimal classification: 8343. Original manuscript recieved by the Institute, November 14, 1928. Presented before New York meeting of the Institute, March 6, 1929. 501 502Beers and Carlson: Developments in Superheterodyne Receivers frequency converting oscillator and detector. The high amplification and great selectivity of this intermediate frequency amplifier largely account for the remarkable performance of the superheterodyne receiver. The ease of obtaining high amplification and selectivity in an intermediate frequency amplifier is chiefly due to the low frequency used and to the fact that thp characteristics of such an amplifier are independent of the broadcast frequency to which the set is tuned. At these lower frequencies by the use of coupled circuits it is possible to obtain the so-called band-pass filter selectivity characteristics. Improved radio -frequency circuits and the use of a higher intermediate frequency have contributed a considerable re -

II ill Fig. 1 duction in hiss and extra responses common to previous superheterodyne receivers. The adjacent channel selectivity and the fidelity of broadcast receivers using a given number of tuned circuits are so related that it is impossible to increase either of these characteristics beyond certain limits without a corresponding sacrifice in the other. A radio receiving set may be designed using four tuned circuits which will have a better adjacent channel selectivity than an eight tuned circuit receiver which is designed for good fidelity. The four tuned circuit set will be lacking in fidelity due to the attenuation of the side bands, and will also have less selectivity for stations separated by 30 kc or more.If the four tuned circuit set is made to equal the fidelity of the eight tuned circuit set, it will be inferior in selectivity in every way. The use of broader radio -frequency and intermediate frequency circuits in connection with a negatively biased detector Beers and Carlson: Developments in Superheterodyne Receivers503 capable of high voltage outputs and a single stage of audio - frequency amplification has contributed a major improvement in the fidelity of the modern superheterodyne receiver. The incorporation of a practical automatic volume control has eliminated the necessity of repeated adjustment of the volume control both when tuning from distant to local stations or vita versa, and when receiving stations whose field strength is varying periodically.

RADIO -FREQUENCY AMPLIFIER The selectivity of a well designed superheterodyne receiver for eliminating local stations is chiefly determined by the selec-

Fig. 2

tivity characteristics of the intermediate frequency amplifier. There are, however, a few exceptions; for example, if no radio - frequency selectivity is employed, two signals of equal strength differing in frequency by twice the intermediate frequency will be received with equal intensity with the heterodyne oscillator tuned midway between them. In addition to this objectionable response, harmonics of the oscillator heterodyning undesired signals are likely to cause interference.An input circuit tuned to the signal it is desired to receive reduces the possibility of interference to some extent, but in the vicinity of powerful broadcasting stations more than one tuned circuit is necessary. The usual type of tuned radio -frequency transformer consisting of a few turns in the plate circuit of the amplifier tube coupled to the tuned secondary gives a selectivity characteristic which is much sharper at the low -frequency end of the broadcast range than at the high. In fact, it is usually so sharp that two 504Beers and Carlson: Developments in Superheterodyne Receivers or three stages of this type of amplification results in considerable reduction of the high modulation frequencies. A transformer of this type designed for good selectivity at the high -frequency end of the range will usually be lacking in amplification at the low -frequency end of the range. A new radio -frequency system has been developed which overcomes most of the objections to the type just described.It provides much more uniform selectivity and amplification without resorting to some mechanical means of varying the coupling with frequency.The only difference between the new transformer and the old is in the type of primary used. Where the primary for the old type of radio -frequency transformer was

MI a a

40 10 10 at o a to to KILOCYCLES OFF lir-5~CE Fig. 3 wound with a small number of turns and was resonant to a frequency above the high -frequency end of the broadcast range, the improved transformer has a large number of turns on the primary, making it resonant to a frequency below the low - frequency end of the range.The old primary increased the effective secondary resistance at the higher frequencies where it was normally too great for good selectivity. The new primary increases the effective secondary resistance at the low -frequency end of the range where it is normally too low for high fidelity. Since the primary is resonant at a frequency below the broadcast range, the amplification is increased at the low frequencies and reduced somewhat at the high, making it uniform over the range. In order to realize the normal amplification of high inductance primary transformers, some meant must be used to compensate Beers and Carlson: Developments in Superheterodyne Receivers505 for the effect of the grid to plate capacity.This is due to the primary being tuned to a lower frequency than the secondary, thus giving the plate circuit capacitive reactance. The voltage fed back through the tube capacity is therefore of such a phase as to oppose the applied grid voltage and will reduce this voltage to a fraction of its normal value.Either of the methods shown in Fig. 1 may be used to overcome this effect.If the feed -back capacity is made too large, it is possible to make the circuit oscillate. Fig. 2 shows one form of a transformer of the high inductance primary type. Curves A, B, and C, Fig. 3, are resonance curves taken at 600 kc showing the effect of the two types of primaries MIIMID7411.1=MMI=1111 M=111==IMK1/11/111111.

mmonimingir.IAMEINMINMW.11MiNUMM/AMMIMMEIMMI 111WAIVIIIIMMININI11101.1111WM IMIIIIMMI111111111=111 IMMMEMIMMEMI

420 ao "up..".'2PESO 20 40 40 00 200 °Cl rJCYCL oFF /ONCE Fig. 4 on the tuned secondary. Curve A is the resonance curve of the secondary alone.Curve B shows this secondary coupled to a high inductance primary, and Curve C shows the same secondary coupled to a low inductance primary. Curves A, B, and C, Fig. 4, are similar curves taken at 1400 kc.Fig. 5 shows the amplification curve for a complete stage of radio -frequency amplification using a high inductance primary transformer and a UY-227 tube. INTERMEDIATE -FREQUENCY AMPLIFIER In past superheterodyne receivers the intermediate frequency has usually been in the neighborhood of 40 or 50 kc. This choice resulted from the ease of obtaining a stable amplifier for a frequency in this region having the necessary amplification and the desired selectivity characteristics. Now, however, 506Beers and Carlson: Developments in Superheterodyne Receivers another important factor must be considered in the choice ofan intermediate frequency. From Fig. 6, showing the selectivity of the two tuned radio- frequency circuits, it is noted that the higher the intermediate

ado 700 660 000 ow Ado atop 400 400 FREQUENCY IV 'mom/is Fig. 5 frequency used the less the possibility of encountering interference from stations separated by twice the intermediate frequency. From the curve it will be seen that this interference for a 40-kc amplifier would be 350 times that for a 400-kc amplifier and 60 times that for a 200-kc amplifier.

JO mminivommo1.

ZF r / FREQUENCY IN KILOCYCLES Fig. 6 An intermediate frequency of 180 kc was finally chosen as the best compromise between amplification, stability, selectivity, and undesired responses. With an intermediate frequency of 180 kc and the oscillator tuned to a higher frequency than the radio -frequency circuits, Beers and Carlson: Developments in Superheterodyne Receivers507 it will be seen that broadcasting stations separated by twice the intermediate frequency will not produce interference at frequencies above 1140 kc. The selectivity of the radio -frequency system at the high frequencies is therefore not as importantin this receiver as at the low frequencies. Both the primary and the secondary of the three transformers used in the 180-kc amplifier are tuned, and the two windings are so coupled as to give a broad top resonance characteristic.Curve A, Fig. 7, is for a single stage of the inter -

FEEQUEkGY IN K/LOCYCLE5 Fig. 7 mediate -frequency amplifier as shown in circuit C, having the proper magnetic coupling between theprimary and secondary Curve B shows the combined characteristic of the primary and secondary used as separate tuned circuits with vacuum -tube coupling between them. Such a characteristic would be obtained if these individual circuits were used separately in a cascade amplifier.These two curves indicate the improvement that can be obtained by the use of coupled circuits, provided the losses in the individual circuits can be kept reasonably low. The overall resonance curve for the intermediate -frequency amplifier is shown in Fig. 8. The approximately ideal band-pass filter characteristics of this amplifier should be noted, the band 508Beers and Cartoon: Developments in'Superheterodyne Receivers

width at 50 per cent peakamplitude being 16 kc whileat 1 per cent peak amplitude it is only40 kc. The three transformers are each mounted in individualmetal containers whichserve both to protect the transformer shield it electrically. The and primary is tuned bya compact fixed condenser, and the secondaryby a small adjustable which permits accurate condenser alignment of the threetransformers. The balancing, primary, and secondary tuningcondensers are mounted on a small piece ofbakelite to whichare also riveted

/00 00 /A

40 10

FlfrQUENCY IN KILOCYCLES Fig. 8

the two brackets forsupporting the transformer illustrates the windings. Fig. 9 manner in which these condensers,the transformer windings, and theterminals are mounted.A transformer completely assembled readyto mount on the chassis in Fig. 10. is shown

AUDIO -FREQUENCY SYSTEM The audio -frequency detector in radio receivingsetsis responsible forsome distortion and contributesto such disturbances as microphonichowl and a.c. hum. These objectionablefeatures are considerably other performance reduced, and advantages are gained byemploying a plate circuit detector and byusing but one audiostage. Beers and Carlson: Developments in Superheterodyne Receivers509

The relative merits of the conventional audio system consisting of a grid circuit detector and two audio stages as compared with a combined radio and audio system employing a radio -

Fig. 9 frequency stage, plate circuit detector, and one audio stage are as follows: Changing from a grid circuit to a plate circuit detector in a given receiver results in a sacrifice in sensitivity. This reduction in sensitivity can be overcome by increasing the radio -frequency

Fig. 10 amplification and reducing the audio -frequency amplification. In general, it can be said that substituting a radio -frequency stage of equal amplification for an audio stage will result in no loss of sensitivity from substituting a negatively biased detector for a grid leak and condenser detector. 510Beers and Carlson: Developments in Superheterodyne Receivers

With improved audio response, particularly at low frequencies, when two audio stages are used,it has become in- creasingly difficult to avoid a.c. hum from the power supply, and audio howls due to the microphonic action of tubes. The blasting and breaking up of the sound output when tuning through a local station is ordinarily due to overloading the output tube. This disturbance can be reduced if the circuits are so designed as to allow the detector and output tube to overload at the same time.

4it RP PEAK WErAcc CV cerEcroo GRID Fig. 11 The time lag distortion associated with detectors employing grid leak and condenser combinations is notpresent when plate circuit detection is used. The curve in Fig. 11 shows the relation between the peak voltage on the grid of a UY-227 tube functioningas a negatively biased detector and the peak audio voltageon the grid of the audio amplifier tube.The plate potential on the UY-227 tube is 180 volts and the bias 25 volts.The carrier is modulated 15 per cent at 100 cycles. The reduction in output foran input in excess of 30 volts is due to the grid current loadon the tuned circuit, through which the voltage is applied to the detector grid. BeersandCarlson: Developnunds in Superheterodyne Receivers511

AUTOMATIC VOLUME CONTROL A receiving set equipped with a practical automatic volume control has several distinct advantages over the set which lacks this equipment. One of the chief advantages is the ability to change the tuning of the receiver without making any change in the volume control setting.Once the volume control has been set for the desired sound output, either distant or local stations may be tuned in without any further volume control adjustment.If the distant station provides sufficient field strength (depending

AAW04ETEC,O4

" T-RY Fig. 12 on the maximum sensitivity of the receiver) itwill produce practically the same sound output as the local station.It is assumed that the per cent modulation used by both stations is nearly the same. The ability of an automatic volume control to limit the sound output to any desired value is of particular importance in light of the present trend towards the use of output tubes capable of handling considerable power. While automatic volume controls have sometimes been called fading eliminators, this term is, of course, erroneous. An effective automatic volume control, however, does automatically adjust the sensitivity of the receiver while the field strength from a station is varying so as to maintain a constant output 512Beers and Carlson: Developments in Superheterodyne Receivers

There are numerous occasions whena powerful distant station has sufficient field strength atany time during an evening to give a satisfactory loudspeaker signal, but the field strength varies through such a wide range andso frequently that the program cannot be received satisfactorily without continuous adjustment of the volume control.Under such conditions an automatic volume control will function satisfactorily. The only indication that the user will have that the signal is fadingwill be an increase in the ground noise when the signal dropsto such 1111111111MEMMEMMEMMEMEM MMOMMEMMEMMEMEMMEM MENUMMEMMINIMMEMINIM MMUSEMMEMMEMMEMEE MMWMWOMMINIMMEMMMEMM MIME. Immommommommm MIRIMMIaMMEMMEMOIMME MIAMMEMMENSMEMEMEME IIMMEMMEMMOMMEMMEMM MAMMEMEMMINNMEMMM EMMENEMMEMMEMMEMMI UMMEMINIMMEMMININIMM IF40/0 FREQLCNCY MICE0 MIL rspem.5)4rFnesrer aeo Fig. 13 a value that nearly the maximum sensitivity of the receiver is required to produce the desired sound output. There are some types of fading thatare accompanied by a distortion of the audio modulation, and in suchcases an automatic volume control will be of little valueas far as compensation for this type of fading is concerned. The chief objections to past automatic volume control systems have been the number of adjustments required and the use of separate voltage supplies for certain parts of the circuit. To obtain sufficient control some systems also require additional amplification either for the a.c. voltageon the detector grid or for the bias voltage before it is impressedon the amplifier grids. This latter objection is overcome by theuse of but one stage of audio amplification and the corresponding increase in the grid swing on the detector tube.Fig. 12 shows a schematic diagram for an arrangement which overcomes the other objections. The grid of the volume control tube is connected in parallel througha Beers and Carlson:Developments in Superheterodyne Receivers 518 coupling condenser to the grid of the second detector. The voltage drop across a resistor in the plate circuit of this tube is used as additional negative bias on the amplifier tubes and thus reduces the sensitivity of the receiver. The values of this plate circuit resistor and its shunting capacity are so chosen as to give the combination a time constant sufficiently low to prevent the smoothing out of the lower modulation frequencies and still high enough to prevent its action from being sluggish. The bias on the volume control tube is normally adjusted to the point where, when no signal is being received, the tube

MLO ion TOO WO RD MOO MOO MO QM FREQUENCYIVKILOCYCLES Fig. 14 draws no current.Then when the grid swing on the second detector exceeds a certain value, the effective bias on the control tube is reduced and its plate current increased.This increase in plate current causes a corresponding increase in the bias voltage applied to the radio -frequency and intermediate frequency amplifier tubes with a resultant reduction in the sensitivity of the receiver, thus tending to maintain a uniform grid swing on the second detector. By means of a manual control the negative bias on the volume control tube can be increased permitting a larger grid swing on the detector tube before the automatic volume control tube takes effect. In this manner any desired audio output can be obtained. The variation of the bias on the volume control tube as the manual adjustment for the output level also compensates for the manufacturing variations in cutoff of the tubes as automatic volume control tubes. Curves A and B in Fig. 13 show the effectiveness of the automatic 514Beers and Carlson: Developments in Superheterodyne Receivers

volume control for two output levels. Curve C shows the variation in output without the automatic volume control. The use of an automatic volume control in a radio receiver presents some problems which are not encountered when this equipment is not used.Due to the automatic volume control maintaining constant audio output, it is difficult to tune the receiver accurately unless some resonance indicating device is provided. A meter in the plate circuit of either the control tube or the tubes controlled provides a satisfactory resonance indicator.

(

0100Vga. AZ

a CYCLES C PEffSOVANCE Fig. 15

In localities where there is considerable interference due to power line leakage, etc., the noise encountered when tuning between stations with a set equipped with automatic volume control is objectionable unless a sensitivity control is provided. This is due to the fact that when no signal is being received the receiver is automatically adjusted for maximum sensitivity, and the time constant of the automatic volume control circuit is such that it does not limit the sbarp impulses of such interference. The sensitivity control may be a voltage divider which varies the normal bias on the amplifier tubes, or in receivers where an untuned input circuit is employed a voltage divider may be used to vary the signal potential on the grid of the input tube. Beers and Carlson: Developments in Superheterodyne Receivers515

By means of the sensitivity control, the maximum sensitivity of the receiver can be reduced to such a value that the noise encountered in tuning between stations is not objectionable. This sensitivity control will in no way destroy the effectiveness of the automatic volume control.

OVERALL PERFORMANCE The major electrical elements of a modern superheterodyne receiver have been briefly discussed.The performance char -

111 8 Af oelavo IVCYCLE5 PERVCON, Fig. 16 acteristics of a complete receiver embodying these elements are as follows: Fig. 14 is the sensitivity curve of a receiver consisting of two radio -frequency stages, heterodyne detector and oscillator, two intermediate -frequency stages, high output detector, and single audio stage. The receiver employs UY-227 tubes throughout except for the power output tube. The three curves in Fig. 15 show the selectivity of the receiver at three frequencies in the broadcast range. The fidelity characteristic of the complete receiver is shown in Fig. 16. The curves show the high uniform degree of selectivity obtained by the use of the eight tuned circuits while still retaining unusual fidelity characteristics. Proceedings of the Institute of Radio Engineers Volume 17, Number 3 March, 1929

AN EXTENSION OF THE METHOD OF MEASURING INDUCTANCES AND CAPACITIES' BY SYLVAN HARRIS (Roister Radio Corp., Newark, N. J.) Summary-The substitution method commonly employed for measuring small capacities is shown to be a special case of a more general principle. As other special cases of this principle methods are presented for simultaneously measuring inductance and capacity when joined in series and when joined in parallel.The cases discussed indicate the method of application of the general principle to any type of measuring or measured circuit. THE method commonly employed for measuring the capacity of condensers, in which the unknown condenser is connected in parallel with a calibrated condenser in a resonant circuit, is a special case of the more general method described in this paper.The several special cases of the more general method are of interest in that they furnish simple and convenient ways in which to measure circuit elements when

OSCILLATOR 0

Fig.1 connected together in circuit. The general principle of the substitution method can be explained by means of Figs. 1 and 2.A series circuit including a current indicator (e.g., a wavemeter) is coupled loosely to an oscillator.Resonance is obtained by adjusting x2 so that = (1) The coupling between x, and the oscillator is assumed to be so loose that the reaction between the circuits is negligible.The unknown reactance x3 is then connected in parallel with x2, and the latter is then readjusted so that resonance is again obtained. Then Xas xl (2) X3- F X2 * Dewey decimal classification: R230. Original manuscript received by the Institute, November 19, 1928. 516 Harris: Measuring Inductances and Capacities 517

Equating (1) and (2), and solving for the unknown x3:

X2 .1,2 X3 = (3) x2 -X2 This is the general condition for all such measurements, and from it the particular cases can be derived.The wavemeter is the most convenient instrument to use in many of these measurements, so that we may put j = jCOL, XI= X2' (4) WC 2 coC 2' and (3) becomes X3 - (5) W(C21- C2) Case 1.If x3 is a pure capacitive reactance, then equation (5) becomes C3= CY-CS' (6) Case 2.If x3 is a pure inductive reactance, it becomes

1 L3=02(C21_Cf2) (7) Case 3.If x3 consists of an inductance in series with a capacity, as for instance an antenna circuit, then x3 = j(cuL3- licoC3) and (5) reduces to C3+ (C21- C2) L3 = (8a) w2Ca(C2' - Cs) Or, (Cs' - C2) C3 = (8b) w2L3(C2' - -1 In order to obtain the values of L3 and C3 it, is necessary to make two measurements at different frequencies. Let these frequencies be f, and /2, and the corresponding values of (C21- C2) be C' and C". Then (8a) may be written C'+C3C"+C3 L3--col2C,C30)22C"C3 (L2 )2 and solving for C3: (0,2-,022)r,fr,/, fi C3- (9) f2 y 1 1 C' C" 518 Harris: Measuring Inductances and Capacities

Similarly, starting with (8b) and solving for L3:

C,' 1 = gic:" (10) 47,2(f22_fi2)( cien(c022-(012) \c"-c") The quantities C' and C" may be termed apparent capacities. For example, in determining the constants of an antenna, it is regarded as a simple capacity, and its apparent values at two different frequencies are measured as in Case 1.These values,

OSULLATOR O

Fig. 2 substituted in (9) and (10), yield the true capacity and inductance. In applying the method to circuits containing coils which have appreciable capacity, the two frequencies should not differ very widely, as the apparent inductance of the coil changes with frequency. Case 4.If xa is an inductance paralleled by a capacity, then

COL3 x3 -j 1 -w2L3C3 Following the same procedure as in Case 3, we obtain:

cox:, _0)22c/i- - C3- co22 -012 f2)2_

0,22_0,0 f22_ f12

t 3 1 = (12) -co 12022(C 1 _ C") 472fi2f22(C1 _ Ci i)

Discussion R. R. Becher': It seems to the writer that several precautions are necessary in applying the measurement methodsoutlined in this paper. It has been his experience that the procedure taken up in Case 3 and the ' Decatur Manufacturing Co., Inc., Brooklyn, N. Y. Discussion on Harris Paper 519 use of formulas (9) and (10) might lead to erroneous results in some cases. This is due to the nature of the circuit resulting when x3 is replaced by a series arrangement of inductance and capacitance. Such a circuit is resonant to two frequencies, one greater than the natural resonance frequency of the L3C3 path and the other less than this frequency. This condition has been examined= and used in several cases by the writer and a number of formulas were developed to cover the several variations of conditions found in practice. The following is illustrative of the circuit relations:

f (zizga)_VLIC2-FL3C3-1-LIC3± V(LiC:-F LaCz ±LIC3)2 4L1L2C2C3 fx3 2LaC3 This is the ratio of the frequency of x3xix2 paths to the resonant frequency of x3. It is the condition that permits this circuit to be tuned to a wavelength shorter than the natural of x3 that makes the circuit so useful.If x3 is an antenna system lower waves can be tuned to than can be obtained with a condenser in series with the antenna circuit. Used in a measurement circuit it is evident that the measurement frequencies must be selected with care, as it will be impossible to obtain a satisfactory adjustment at a frequency near that at which the x1 path becomes resonant. It can be shown that it is necessary to consider the value of z, (that is L1) in any computations. In some cases it may be desirable to determine the ratio of frequency of x1x2xs Resonant frequency of x1x2 instead of that given by the above formula.In this case the values obtained for either the low or the higher frequency must be multiplied by the factor /1,3C': "Y LI C2 which is equivalent to changing the denominator to 2L,C2.It is possible to convert this formula to a more convenient form for use in measuring work. In Case 4 in the above paper a condition may be found where the resonant frequency of x3 may be near or equal to the oscillator frequency. In this case there will be little or no change in the tuning of the x1x3 circuit after the addition of x3.It seems that other peculiar conditions might be encountered if the two measuring frequencies happened to be on opposite sides of the frequency of x3. Quite aside from the above comments it seems that the accuracy of the above method is determined by the ratio of f2/l1.However, mention is made that the two frequencies should not differ very much. If the distributed capacity of the wavemeter pickup coil is measured and the value added to that of the condenser and the precaution taken to use as large a value of C2 and as small a value of pick up inductance the error is 2 "Prepared Radio Measurements," pages 89 to 91; and Wireless Age, p. 39, April, 1919. 520 Discussion on Harris Paper somewhat lessened and it might be possible to use a larger frequency ratio. Since we are mainly concerned with the difference between two readings of C, the effect wavemeter coil capacity which is effectively in parallel with the condenser may not introduce as great an error in results as might be found with the use of two measuring frequencies differing but little. The effect of the distributed capacity of the coil under measurement is rather hard to predict, however. Proceedings of the Institute of Radio Engineers Volume 17, Number 3 March, 1929

APPARENT EQUALITY OF LOUDSPEAKER OUTPUT AT VARIOUS FREQUENCIES' BY L. G. HECTOR AND H. N. KOZANOWSKI (Department of Physics, The University of Buffalo, Buffalo, New York) Summary-A type of alternation phonometer has been developed which permits rapid switching of power at two frequencies to the same loudspeaker without the distracting effect of the transients that would result from ordinary types of commutation.This result is obtained by the use of rotating condensers to provide variable capacitative reactance in the input circuit of the power amplifier that operates the speaker.The power consumed by the loudspeaker is measured with a specially constructed wattmeter of the electrodynamometer type and the output of the loudspeaker is measured by means of the torques produced on a Rayleigh disc.With the aid of the alternation phonometer and an additional capacitative reactance, the observer is able to adjust the power input to the loudspeaker until two tones of different frequencies appear to have the same intensity. The purpose of the research was to develop a method for the comparison of loudspeaker efficiency at various frequencies that could be used in ordinary laboratories with limited equipment. PREVIOUS WORK 11HAT the sensitivity of the human ear is a function of the frequency of the tone has long been assumed, and quantitative results on this sensitivity for minimum audi- tion have been obtained by Fletcher and Wegel.' They employed a telephone receiver clamped over the ear of the observer as a sound generator. For tones of greater intensity, Dayton C. Miller' employed a set of organ pipes which when driven under certain standard conditions of air pressure gave tones which in the opinion of the builder (an expert pipe organ man) gave tones of equal apparent intensity. B. A. Kingsbury' compared tones of different pitch for the sake of determining the actual pressure on the ear drum for equal apparent intensity at various pitches.He employed a single receiver clamped over the ear after the manner of Fletcher and Wegel. The observer varied the intensity of one tone by means Dewey decimal classification: 8376.3. Original manuscript received by the Institute, November 20, 1928.Presented before meeting of the Buffalo section of the Institute, January 17, 1929. 1 Fletcher and Wegel, "Frequency Sensitivity of Normal Ears " Proc. Nat. Acad. Sci. 8, 5, 1922. 2 Dayton C. Miller, "The Science of Musical Sounds" Macmillan Co. 8 B. A. Kingsbury, "A Direct Comparison of the Loudness of Pure Tones," Phys. Rev., 29, 588; April, 1927. 521 522Hector and Kozanowski: Loudspeaker Output at Various Frequencies of an attenuator, and he shifted his attention from one tone to the other by manually operating a double pole double throw switch so that he listened first to one tone and then to the other. Previously, Donald MacKenzie had pointed out certain defects in this method of alternating the tones due to the "decay of sensation" on the part of the observer. He therefore devised an instrument called an "alternation phonometer," which bears some relation to the flicker photometer of optics.Power from two oscillators was alternately impressed on a sound generator by means of a polarized telegraph relay driven from another relay at controllable speeds. The speed found to be most satisfactory was approximately 12 per second in the frequency region 100 to 4000. That is, each source was turned on and off 12 times per second. Only about 0.002 second elapsed between the opening of one contact and the closing of the other. A ther-

113 s+_13 Fig. 1-The Variable Reactance Alternator with Associated Apparatus. mophone held tightly to the ear was used as the sound generator for part of this work, and an electro-magnetic receiver similarly placed was used for the part involving intensities above the capacity of the thermophone.

PURPOSE AND GENERAL METHOD The present paper is an account of work done by the authors in attempting to apply the general principle of the alternation phonometer to a method for the determination of the apparent efficiencies of loudspeakers at various frequencies. The general method was to adjust one oscillator to give a steady tone at an arbitrarily chosen intensity. The observer then adjusted the intensity of the other tone until he obtained apparent equality. The two tones were rapidly alternated by a new type of alternation phonometer. The electrical power input to the loudspeaker was then determined for each tone by means of a specially con- Donald MacKenzie, "The Relative Sensitivity of the Ear," Phys. Rev., 20, No. 4; October, 1922. Hector and Hozanowski: Loudspeaker Output at Various Frequencies523 strutted electrodynamometer calibrated at each frequency used. The actual sound intensity at the position of the observer's head was determined from Rayleigh disc measurements.

THE ALTERNATION PHONOMETER In the place of MacKenzie's relay, the authors attempted to use a rotating commutator, but found that with the amplification employed between the commutator and the loudspeaker, the transients introduced by the sudden starting and stopping

Fig. 2-The Variable Reactance Alternation Phonometer. of the power from the oscillators were fatiguing to the observer and seriously interfered with his ability to compare the intensities of the tones. An arrangement of variable capacitative impedance in the input of the amplifier was then tried, and it proved to be much more satisfactory. In the diagram of Fig. 1, Cl and C2 are variable air condensers of 0.0005 pf maximum capacity. They are mounted with their rotors on a common shaft so that one condenser is at minimum capacity when the other is at maximum value. Fig. 2 shows two condensers for Cy.These are electrically insulated and are arranged so that they may be used either independently or in series or in parallel.It was found necessary to use only one of them independently in the work reported in this paper.The shaft on which the rotors are mounted has a pulley on one end 524Heotor and Hozanowski: Loudspeaker Output at Various Frequencies and is belt driven by a variable speed d.c. motor. The speeds most commonly used gave 8 or 9 alternations per second. That is, each tone was heard that number of times per second.

INTENSITY CONTROL In the diagram of Fig. 1, C4 is a variable condenser which enables the operator to make an arbitrary setting for the maximum value of the intensity of the tone to be used as a basis for comparison.C3 is used to vary the intensity of the tone

-r;(2.01-A) -c" (20 Cs 11111111 nil

(2.00) 2 11111III T Fig. 3-Oscillator with Conductively Coupled Amplifier. being compared.It is operated by a long shaft running into the sound compartment so that the observer may readily operate it. Ca and Cg were large condensers (4 /If each) used merely to keep the platepotentials off the control condensers C3 and Cy,. T3 and T4 were amplifiying tubes which were in turn fed by oscillators. OSCILLATORS AND AMPLIFIERS The oscillators were of the Hartley type. Air core coils were employed and tuning was accomplished by means of step variable condensers. Each oscillator was conductively coupled to a one -stage amplifier. The arrangement of one of these oscillator - amplifier combinations is shown in Fig. 3. The controlled output from the rotating condensers C1 and C2 was fed into a resistance coupled amplifier of conventional design. The first tube of this amplifier is shown in Fig. 1 and the entire arrangement in Fig. 4.The amplifier is built into a brass - lined box with brass partitions between the separate stages. Separate filament batteries are used for each oscillator and for the resistance coupled amplifier.But the circuits are so arranged that although separate B and C batteries are used also, the negative sides of all the B batteries used on the amplifiers are electrically connected. This is of course necessary for the operation of the variable reactance in the grid circuit of the amplifier input. Hector and Kozanowski: Loudspeaker Output at Various Frequencies 525

LOUDSPEAKER The loudspeaker used in these experiments was an exponen- tial horn driven by a unit of the Baldwin type. The horn was octagonal in cross section and was built of bass wood. The length

B 2C; Fig. 4-Resistance Coupled Amplifier.The coupling condensers are 1µf each.The filaments are in parallel with a common rheostat. Both grid and plate resistances aremegohm each. along the axis was 8 ft.The open end had a cross section of 400 sq. in. and the small end a cross section of approximately 0.3 sq. in. The equation for the shape of the horn is therefore given by A=,(0.075.-1.181) where A is the area of cross section in square inches, x is the distance in inches measured along the axis from the smaller end, and e is the base of the natural system of logarithms.

Fig. 5-The Loudspeaker and the Sound Compartment.

SOUND COMPARTMENT The mouth of the loudspeaker opened into the end of a 9 -ft. x 4-ft.x 8 -ft. compartment as shown in Fig. 5.The box was 526 Hector and Kozanowski: Loudspeaker Output at Various Frequencies built of M in. celotex mounted as shown on a 2-in.x 4 -in. lumber frame. It was lined on the inside with 2 layers of M -in. hair felt. On the under side of the box, at the end opposite the loudspeaker, a little of the celotex was cut away, and the hair felt folded back to permit the observer to stand upright on the floor of the laboratory with his head approximately along the mid -axis of the compartment and near its rear end wall.The long shaft from the volume control condenser, C3, entered the compartment near this end, so that the observer could easily adjust the tone intensity. RAYLEIGH DISC MEASUREMENTS When Rayleigh disc measurements were to be made the hair felt was closed and the celotex replaced.The Rayleigh disc used in these experiments was a piece of mica 8.75 cm in diameter

Fig. 6-Rear End of the Sound Compartment Showing the Window for Making Observations and the Octagonal Coils for Establishing the Restoring Field to the Magnet. suspended from one edge by a thread of unspun silk.On the mica was mounted a small galvanometer mirror which received and reflected light through a narrow glass window in the rear end of the box. (See Fig. 6.) Since the zero position for the disc Hector and Kozanowski: Loudspeaker Output at Various Frequencies 527 was 45 deg. to the axis of the box, it will be seen that this window was not directly behind the disc, but rather near one edge of the end wall. The disc also carried a tiny magnet (1.49 cm in length and 0.1389 gram weight) mounted at 45 deg. to the plane of the disc.With the major axis of the sound compartment running north and south, the disc was then in its zero position (45 deg. to the axis of the compartment) when the small magnet was parallel to the earth's magnetic field (H.). Essentially a Rayleigh disc depends for its operation on the fact that when an object finds itself in a field of uniform flow it tends to set itself with its greatest dimension cross -wise to the direction of flow. A quantitative development' for the case of a round thin disc located in a field of alternating flow where the alternations are simple harmonic gives Torque = 8/3ir2a2n2pC3 sin20 where a is the amplitude of the vibration n the frequency p the density of the medium C the radius of the disc 0 the angle between the normal to the disc and the direction of flow. Now since the mean square velocity in simple harmonic motion is given by 272a2n2, and since the energy in a sound wave is proportional to the mean square velocity of the vibrating particles it follows that the torque on the Rayleigh disc is proportional to the intensity of the sound at the point considered. Obviously this torque is a maximum when the angle 0 equals 45 deg. The sound waves striking the disc tended to rotate it and the attached magnet.As the magnet moved out of the earth's magnetic field, the magnetic couple due to that field tended to restrict the motion. A second magnetic field, H, at right angles to the earth's field was then established by sending a measured current through the, coils shown in Fig. 6.The couple due to this field also tended to return the magnet (and hence the disc) to its original position. Since the couple due to the earth's field is equal to zero when the magnet is in its original position the final magnetic couple, G, is given by Winklemann, Handbuch der Physik., II.p. 238. 528 Hector and Kozanowski.: Loudspeaker Output at Various Frequencies

G= H M where H is the field due to the current in the coils and M is the magnetic moment of the magnet. But this couple must just balance the torque on the disc due to the sound waves. Hence, since H is proportional to the current in the coils, this current will be directly proportional to the intensity of the sound. Since relative loudspeaker efficiencies were the chief concern in this work, these torques are here expressed in terms of this current in milliamperes.

.1;00 Ia 0.6" 700

boo A

ks 500

oea LOO yo U 300 ti °a 200 4-

100 2-00

100 200300 900500 600700 500 900 1000 POWER nMILLI -WATTS Fig. 7-Sound Intensity Compared to Input to the Loudspeaker at Various Frequencies. To determine zero position for the disc, light sent through the window and back from the mirror mentioned in a preceding paragraph was focused at the source and received on a direct vision screen. Since the operation of the Rayleigh disc requires considerable care, all the sound intensity measurements were made at one time and the data at each frequency plotted against electrical input to the loudspeaker.Fig. 7 shows the calibration curves for the various frequencies used. With the aid of these curves the sound intensity could be determined for each setting of an observer by simply reading the electrical power input.

WATTMETER For the purpose of measuring the power input to the loudspeaker a wattmeter of the electrodynamometer type was constructed. The dynamometer with the front of the case removed Heotor and Hozanowski: Loudspeaker Output at Various Frequencies 529 may be seen in Fig. 8, and a wiring diagram showing internal and external connections of the instrument is shown in Fig. 9.It will be seen that the instrument functions as a simple indicating wattmeter.

Fig. 8-The Electro-Dynamometer with the Front of the Brass Case Removed. The moving coil of the wattmeter consists of 2,100 turns of No. 40 enameled cupron wire wrapped on a fiber form of about 2.0 cm diameter. A resistance of 50,000 ohms is in series with this coil. The coil resistance is 3740 ohms giving a total of 53,740 ohms for the potential circuit of the dynamometer. The stationary coil is divided into two parts with the movable coil between them as shown in the photograph.The stationary coil is made of 1,200 turns of No. 28 enameled copper wire.Its resistance is 37 ohms. 530Hector and Kozanowski: Loudspeaker Output at Various Frequencies

The movable coil has a standard galvanometer suspension of phosphor bronze ribbon above and a coil of phosphor bronze ribbon beneath itself.This coil makes contact through a small copper container which carries oil.Two wire prongs extend downward from the moving coil into this oil and serve to damp the oscillations of the coil.A galvanometer mirror is also mounted directly on the moving coil.

Potential.

S°7re 5 Power Fig. 9-Schematic Diagram of the Wattmeter Showing Connections to the Source of Power and to the Loudspeaker. To avoid any errors that might arise from lack of shielding, frequency variation, energy used by the dynamometer itself, etc., the instrument was calibrated in place at every frequency and throughout the range of power used.Fig. 10 shows the

6 Fig. 10-Schematic Diagram for the Calibration of the Wattmeter. arrangement used for calibrating.Power was supplied at the various frequencies to a pure resistance. The current was measured by means of a Western Electric thermocouple calibrated with direct current and a milliameter. The deflections of the wattmeter are read by means of a telescope and scale placed at a distance of 200 cm. Fig. 11 shows a calibration curve for 25 cycles, 250 cycles, and 1,200cycles. Table I shows the necessary data for the use of the dynamometer at all the frequencies employed. TABLEI SZNOITIVITT WATTLICTZR Frequency Milliwatta per em. 67.0 170 68.5 250 70.4 400 73.0 1.200 77.3 Hector and Hozanowski: Loudspeaker Output at Various Frequencies 531 While the alternation phonometer was operating and an observer was adjusting the intensity of one of the tones the switch K shown in Fig. 9 was turned so as to cut the potential coil out of the circuit and put a conductor of approximately equal resistance in its place. As soon as the observer decided on his adjustment, the rotating condensers were stopped with either C1 or C2 in its maximum position. Energy to. the other condenser was cut off by removing the plate potential from the oscillator -amplifier feeding it.The potential coil of the dynamometer was then placed in the circuit by means of switch K

$15

13 qll 4,0 .4 J58 0z vµ5 w a

100 100300 400500 600 700 cf00 1000 POWER MILLI -WATTS Fig. 11-Calibration Curves for the Wattmeter at Various Frequencies. and the deflection of the dynamometer determined. The condensers were then rotated through 180 deg. and the power input to the loudspeaker from the other source determined in a similar manner.The frequency of the variable source was then altered and the process repeated.It was of course necessary to keep the intensity of the tone used as a basis for comparison as nearly constant as possible.

EXPERIMENTAL RESULTS Fig. 12 shows a single set of observations of sound intensity versus frequency for equal apparent intensity of tones using a frequency of 400 at an electrical input of 755 milliwatts as a base. This data is represented by the crosses. The circles show the average of 3 sets of data taking a frequency of 1,200 at an electrical input of 1055 milliwatts as a base. The squares are the average of 2 sets of data using a frequency of 250 at 332 milli- 532 Hector and Hozanowski: Loudspeaker Output at Various Frequencies watts as a base. The solid line shows the average of 4 sets of data using each tone once as a base. The conformity of the curves for all the observations shown in this figure to the same general shape is an indication of the accuracy of judging equal apparent intensities with the aid of the alternation phonometer described even at rather high loudspeaker intensities. In Fig. 13 we have a record of the same observations as in the solid curve of Fig. 12 except that here we have plotted power input against frequency. This figure is of course not of general interest, since it applies only to the loudspeaker used in these tests.

600.111=11111111111111111MMIMER

3,..11112=M11111111= 11111111111111 .,11111111116111111111111111111

ENIMIIIMMEMIIIMMIIMI01111111.111.111111111 y 5.1111111111111MNIMMMMISSIIIIIIIIMI 0,MMIIIMENIMMUMEN

2111.11111111111111111.1 /00 L50 200 300 MOO SOO NO 4000 4500 2,000 FREQUENCY Fig. 12-Observations of Sound Intensity Versus Frequency for Equal Apparent Intensity. (See text for explanation of marks.)

These curves (Fig. 12) show, as was to be expected from the work of previous investigators, that the ear sensitivity falls off for the lower tones. On the other hand, we see by plotting the ratio of Rayleigh disc torques to electrical input against frequency for eqiial intensity (data for the observations shown in the solid curve of Fig. 12) that the loudspeaker used is more efficient in our lower range than in the upper.(See Fig. 14.) A similar curve plotted for equal actual intensities would show this characteristic still more. Except for a frequency of 170, Fig. 7 indicates that the efficiency of this loudspeaker does not vary with intensity at a given frequency in the region covered by this graph. Hector and Kozanowski: Loudspeaker Output at Various Frequencies 533

SOURCES OF ERROR AND LIMITATIONS OF THE DEVICE The alternation phonometer did not prove as satisfactory as a manual switching of sources when both tones were at approximately the same frequency, due largely to beat notes. The plates of condensers C1 andC2were semicircular.It will be readily seen that three questionable things occur with commutation by means of these condensers. First, at no time is

5 *0,1111MISHIIIIIIIM111111111111111111111 100 130 200 300 1100 .500 700 1,000 1,5002,000 FR E qtAENCY Fig. 13-Power Input Versus Frequency for Equal Apparent Intensity the capacity of either condenser zero, and consequently at all times both tones are present to some extent. Second, the intensity of either tone approaches a maximum and at once starts to recede.It does not stay at constant intensity even approximately for any large percentage of the time. Third, twice during every cycle of the alternator both tones are present simultaneously at the same intensity. Whether or not any or all of these items are seriously objectionable is difficult to say. The second item could probably be rectified by using condenser plates of different shape. It is probable that the absolute values of sound energy as could be determined from the Rayleigh disc torques might be in error, in the first place, due to another torque from radiation pressure. The possibility of error from this cause can be obvi- ated by repeating the measurements with the disc rotated through 90 deg. and the two sets of data averaged.6 It was not thought necessary to take this precaution for the purposes of this experiment. There is also the possibility of error due to reflections of sound waves from the walls of the compartment.Essentially this phenomenon may be thought of as sound waves approaching 6 S. L. Quimby, "On the Experimental Determination of the Viscosity of Vibrating Solids," Phys. Rev., 25, 558, 1925. 534 Hector and Kozanowski: Loudspeaker Output at Various Frequencies the disc from various directions. The result of such reflections may be either to increase or decrease the torque on the disc. Judging from the data of Watson,? it seems unlikely that much more than 25 per cent of the incident energy could be reflected. From the general nature of the Rayleigh disc calibration curves for various frequencies shown in Fig. 7 it would seem that no very prominent standing wave patterns existed in the sound compartment. Had they existed, the disc would of necessity have found itself in different parts of the pattern for different

1.1 Ia 0 OL 11

yy a< (O4

201

100 150 ZOO 300 1003 300 700 1,000 1,3003,000 FREQUENCY Fig. 14-Apparent Relative Efficiency of Loudspeaker. frequencies.These curves do not indicate such an effect. It would seem that the value of the torques on the Rayleigh disc are at least dependable for relative values of intensity. We hope, however, to improve this feature for further measurements. The principal source of error is probably traceable to the use of too small a power tube to feed the loudspeaker. This matter will also be taken care of in future measurements.

UTILITY OF THE ALTERNATION PHONOMETER In some respects this paper is in the nature of a preliminary report. However, its publication at this time seems warranted by the expectation that it describes a method applicable to loudspeaker measurements in laboratories with limited facilities for such measurements. In many cases of practical work with loudspeakers the matter of relative efficiency at various frequencies is of more importance than actual efficiency of the speaker. It is therefore hoped that the work here described may suggest the possibility of making 7 F. R. Watson, Phys. Rev., 7, 125, 1916. Hector and Kozanowski: Loudspeaker Output at Various Frequencies 535 analyses of loudspeakers over a considerable frequencyrange in commercial work.For such work the Rayleigh disc could be omitted and only the wattmeter and the alternation phonometer used. For rough work the sound compartment could be omitted, but for more nearly reproducible results it will be found quite necessary.

ACKNOWLEDGMENT The authors are indebted to Mr. L. C. F. Horle, of the Federal Radio Corporation, for advice and suggestions during theprog- ress of this work. Proceedings of the Institute of Radio Engineers Volwme 17, Number 3 March, 1929

FACSIMILE PICTURE TRANSMISSION* By V. ZWORYKIN (Research Laboratory, Westinghouse Electric and Manufacturing Company, East Pittsburgh, Penna.) Summary-A facsimile picturetransmittingsystemisdescribed. The chief object of the design of this system was to produce a simple, rugged apparatus for practical usage, which would not require the attention of a skilled operator.The system does not require a special preparation of the original, and the receiver records the copy directly on the photographic paper. The usually delicate problem of photo -cell current amplification has been simplified to such an extent that only three stages of resistance coupled amplification suffice between the photo -cell and modulator of the broadcasting station. This was made possible through the design of a very efficient optical system, which supplies to the photo -cell quite enough light reflected from the picture even though only a small incandescent lamp for illumination is used. The synchronizing and framing have also been simplified to such a degree that they do not require any special channels or special amplifiers. Automatic starting devices obviate the use of any complicated scheme of signal dispatch for starting the apparatus.In spite of the simplicity of operation, it is capable of transmitting a 5 in. by 8 in. picture either in black and white or in half -tone in 48 seconds, or a message at the rate of 830 words per minute over short distances. The resulting picture prints are of a quality quite satisfactory for news- paper reproduction and clear facsimile of messages may be made from type- written originals. OPTICAL SYSTEM LL the existing methods of electrical picture transmission can be divided into two classes: one which requires special preparation of the original before it can be transmitted, and the other which can transmit the original directly. The first one includes the electrical contact method, now almost abandoned, which requires the preparation of the original in such a form that the dark and light of the picture give variable electrical resistance when explored by a traveling contact. In another form of transmitter of the same class, as in the Belin system, which is still in commercial use in France, the picture is embossed with a special ink so that it may be reproduced by a microphone in the same way as an electrical pick-up reproduces phonograph records. The method which requires the preparation of a transparent picture either in negative or in positive form belongs also to Dewey decimal classification: R582. Original manuscript received by the Institute, November 24, 1928. Presented before New York meeting of the Institute, January 2, 1929. 536 Zworykin: Facsimile Picture Transmission 537 the same class.' In this case the picture is explored by a sharply defined pencil of light which passes through it and activatesa photo -cell placed behind it.The variation in optical density produces a corresponding variation in absorption of the light, and therefore the photo -cell delivers an electric current varying according to the picture. The present requirement of high-speed transmission, however, rules out all these methods due to the time necessary for preparation of specially treated originals.In this case only one solution remains, and this is the scanning of the original by a pencil of light and the utilization of the light reflected from its surface. The amount of reflected light is directly dependent on the density of the picture or lettering. The specular reflection

PAIPAICILIC AlirOYecreor R,ortioag /Ai I'M - -_0Alirenieg

JoVOCE OI ARRROR lad MT OR ORRIIRR011 MOMS OW! SIGAIll FOR AtROMMC !WPM, Fig. 1-Optical System of the Picture Machine. from the surface of the paper is negligible and equal at all points on the paper, and therefore does not interfere with the reflecting scanning method. However, difficulties arise in the optical part of the problem due to the small amount of light reflected. This necessitates the utmost care in the design of a very efficient optical system. Here again are two possible solutions of the problem. One is the illumination of the original by strong diffused orconcen- trated light and cutting off from the illuminated areaa small, sharply defined spot. The light from this spot is directed intoa photo -cell.In this case the optical efficiency is determined first by the ratio of the size of the scanning spot to that of the total illuminated area, and by the light -gathering power of the optical lens.In general, in spite of all the precautions, the over-all optical efficiency of this method is quite low. a This method is still in general use and, a few years ago, was the only optical method employed. 538 Zworykin: Facsimile Picture Transmission

The second method, which is used in this transmitter, is inverse to the first one. In this case the size of the scanning spot is adjusted to the required dimension and the reflected light is collected. Fig. 1 gives an idea of the arrangement. The source

MOTOR 3

116H7cmcv.piX,

"Mr

Fig. 2-Facsimile Picture Transmitter. of light is focused first on a diaphragm to make the size of the spot independent of the size of the source.The image of the diaphragm, with necessary reduction, is focused on the surface of the picture. The reflected light is gathered by means of the parabolic reflector, the focus of which coincides with the illuminating point. Part of the reflector is cut away in order to pass the light, and the remaining part is brought into close proximity Zworykin: Facsimile Picture Transmission 539 to the surface of the picture. In this case almost all the reflected light is collected and projected as a more or less parallel beam by the reflector. A plane mirror with a small hole for passage of the illuminating spot intercepts the reflected light at 45 degrees and diverts it to the photo -cell. The optical path of the whole system is quite short and the construction is flexible for adaption to almost all kinds of scanning systems. The over-all optical

Fig. 3-Glow Tube. efficiency is many times greater than the best possible solution by the first method. SCANNING ARRANGEMENT In the present article only the intermittent type transmitter, i.e., the type in which it is necessary to stop the machine after every picture for reloading, is described. Although this typeis not very suitable for commercial purposes, it has many advantages for experimental work and also for all kinds of communication where traffic requirements are not very high. In this type, the picture to be sent is in the shape of a 5 in. by 8 in. rectangle and is wrapped around a cylinder which is placed on the shaft as shown in Fig. 2. The shaft is hollow and 540 Zworykin: Facsimile Picture Transmission

has a screw inside whichcan be locked either to the shaft or to the support. The cylinder hasa locking nut, which, by means of a lever, can be raised or lowered, locking the cylinder with the shaft and screw. While thescrew is connected to the shaft, the cylinder rotates with it, but remains stationary inlongitudinal direction. When the picture is ready for transmission,the screw is released from the shaft and locked to thesupport by means of an electromagnetic clutch. This makes the screw stationary in respect to the rotating nut and the cylinder begins to advance,

H 1t z

CV 10

0 2_ 4 6 6 10 12. 14 MILLIAMPERES EXPOSING CURRENT Fig. 4-Fidelity Curve. exposing gradually the whole surface to the scanningspot. The speed of scanning is 56 in. per second, while that of the cylinder feed is 1/64 in. per revolution, so that 52.5sq. in. of the picture are covered per minute.Allowing, as usual, 12 wordsper sq. in., the rate of transmission amounts to 630 words per minute. The 1/64 in. picture feed was found very satisfactoryfor type- written messages and for most other pictures. Care, of course, is taken to move the cylinder with uniform speed in order to avoid distortion of the picture. Mechanically, the receiver is identical with the transmitter. A standard bromide photographic paper 5 in. by 8 in.size is wrapped around the cylinder and the recording is done bymeans of a glow discharge tube. Fig. 3 shows the type of glowtube used for this purpose.It was developed by Mr. Knowles in the Zworykin: Facsimile Picture Transmission 541

Westinghouse Laboratory and uses helium glow for recording. The glow is restricted to the required size by a mask built into the tube. A discharge of approximately 15 milliamperes at 400 volts is sufficient at the present speed to produce very satisfactory blackening on the bromide paper. Fig. 4 shows the fidelity curve for the whole transmitting process. Along the axis of abscissas is plotted the current through the glow tube.The bromide paper is exposed to this glow at working speed and is developed in the usual manner.The "density chart" prepared in this manner is put into the trans -

Fig. 5-Photo Cell. mitter, and the output of the amplifier is plotted along the axis of ordinates. The curve, therefore, represents the true relation between the densities of the original and of the reproduced copy. This relation shows good proportionality from 2 to 12 milliamperes, which ratio usually is maintained for transmission of half tone pictures.For white and black manuscripts it is preferable to sacrifice the proportionality and work further along the characteristic curve in order to produce sharper contrast. For a highly efficient commercial transmitter, the intermittent type is not suitable, due to loss of time required for reloading.Another machine of continuous type has been developed for this purpose. Although the principle and even the optical system remains the same, the relative motion of the paper with respect to the scanning light is changed, andthis 542 Oworykin :FacsimilePicture Transmission

in turn changes the whole appearance of the apparatus. The reflected light is conveyed by plane mirrors along the axis into a photo -cell, which remains stationary.

PHOTO CELL Fig. 5 shows a photo -cell used in the picture transmitter. It is of magnesium -caesium type, filled with argon.It consists of two electrodes, one on the inner surface of the glass bulb and another in the shape of a ring in the center of the bulb. A window is provided in the coating for admittance of the light. The coating is photo -sensitive; i.e., it emits electrons at a rate r.rligirt geHishiiiillianu Boa

1 15111"MilMill V 1111Oil 3 MI mummawnu 1 .moseurailli. . VOLTAGE Fig. 6-Characteristics of Gas -Filled Photo -Electric Cell. proportional to the quantity of light absorbed by the coating. These electrons flow to the anode under the accelerating potential of an outside battery. During this passage they collide with the molecules of the argon, and since their velocity -voltage is higher than the ionizing potential of the argon, ionization occurs. Thus the output of the cell is increased many times without destroying the proportionality between the quantity of light absorbed by the photo -cell and the output of the cell.Fig. 6 shows the relation between the voltage applied to the cell and its output calculated in microamperes per lumen. The second curve on the same figure gives an idea of the safe operation of the cell for various voltages and degrees of illumination.

AMPLIFIER Since the photo -cell, under operating conditions, supplies a current of the order of 1/20 of a microampere for the white Zworykin: Facsimile Picture Transmission 543

portion of the picture, a strong amplification is necessary before the output of the cell can be used for radio transmission. The requirements for the amplifier are quite severe.It should not distort the signals and should not havea tendency to oscillate, which results in distortion of the picture.

TO MOOULATOlt CW TWANoWIT TER Fig. 7-Photo-Cell Amplifier. In actual cases we used screen -grid tubes and the circuit as shown in Fig. 7.Voltage output of the third tube is about 40 volts, which is quite sufficient to operate the modulator of the broadcasting station through a line of considerable length. Fig. 8 shows an oscillograph actually received from the picture signals.

Fig. 8-Envelope of Signal Current. RECEIVER AND AMPLIFIER For the reception of the picture signals, the radio set may be a standard receiver. The amplification, in the case of weak signals, is preferably at radio frequency in so far as possible, in order to reduce distortion. Transformer -coupled audio -frequency amplification, however, gives very good results if the gain is 544 Zworykin: Facsimile Picture Transmission fairly uniform between 2000 and 4000 cycles. To date, a standard R.C.A. short-wave receiver has been used for all work.This employs a stage of radio -frequency amplification with a screen - grid tube, detector, and two audio stages.

CONTROL OF GLOW TUBE For the control of the glow tube which exposes the photographic paper, a vacuum tube is used. The glow tube is connected in series with the plate voltage for the vacuum tube. Fluctuations in voltage upon the grid due to the picture signal produce corresponding variations in the glow tube current.If the signal as it comes from the audio amplifier were applied

RECEIVING SET AND AMPLIFIER

GLOW TUBE

ROTARY SNITCH ON CYLINDER SHAFT Fig. 9-Glow Tube Control Circuit. directly to the grid of the control tube, a negative picture would result. That this is true can be verified by following the steps in the transmission of the picture. When the light is reflected from a white area in the original picture, a relatively large amount reaches the photo -cell. The corresponding photo -cell current is amplified, producing a loud signal.At the receiving end this signal would increase the average plate current of a tube working on the lower bend of the characteristic curve. Such an increase would augment the light from the glow tube, darkening the photographic paper instead of making it lighter. Unless the picture is to be recorded upon a film, and sub- sequent prints are to be made, it is necessary to reverse the process. This reversal might take place at the transmitter, but is undesirable for pictures which are largely white, as printing, for example. Bursts of static would be recorded as black spots Zworykin: Facsimile Picture Transmission 545 on the white background. On the other hand, if the reversal occurs at the receiver, these disturbances tend only to make the white whiter. Hence, reversal at the receiver is used. The circuit employed for the control of the glow tube is shown in Fig. 9. Voltage from the receiving set is applied to the push-pull detector, using UX-112 tubes. These are so biased as to give practically zero plate current in the absence of signal. Any voltage supplied causes, on either the positive or negative half of the cycle, a voltage drop across the plate resistor.This

Fig. 10-Synchronizing System. voltage drop is impressed on the grid of the control tube, decreasing the glow tube current in the case of a strong picture signal. SYNCHRONIZING The problem of synchronizing transmitter and receiver is of great importance, particularly for high speed transmission. The plan of broadcasting a standard frequency by a series of stations scattered throughout the world is one of considerable merit, but has not yet been adopted. The use of voltage from inter- connected power lines has been proposed, but is impractical in the general case, since the phase relations between ends of the system are too variable to permit high speed transmission. It is necessary, then, to do one of three things: (1) provide independent, accurate sources of frequency for transmitter and receiver, or (2) send a synchronizing signal continuously to the receiver, or (3)correct periodically a less accurate source of 546 Avorykin: Facsimile Picture Transmission

frequency at the receiver by an impulse from the transmitter. The third method is the one used. The source of frequency at both transmitter and receiver is a 70 -cycle tuning fork in a constant temperature box.These forks are so adjusted that there is but one beat between them in 20 seconds or more; this condition is relatively easy to maintain. The fork at the receiving machine is then corrected every revolution of the picture cylinder (every seventh of a second) by an impulse of about one-half cycle duration. This impulse is transmitted over the same channel as the picture, but on the margin of the paper to avoid interference with picture signals.

RECEIVER AND TO GLOW TUBE AMPLIFIER CIRCUIT

TUNED RELAY

GRIDGLOH TUBE

LOCK -IN RELAY

TO MAGNETIC CLUTCH

Fig. 11-Automatic Starting Arrangement. Having obtained the standard frequency, the next step is to use it most advantageously in the control of the motors.To amplify a small amount of energy to such a degree that it could supply the full load of the machine would be wasteful.It is common practice at the present time to use two motors on the same shaft-one to furnish most of the torque, and one to keep the speed constant.In the present arrangement, the two machines are combined into one, similar to a rotary converter. Voltage from the tuning fork is amplified, using two UX-250 tubes in the final stage. The power from these tubes is applied to the motor slip rings, while most of the energy comes from the direct -current source.Fig. 10 shows the schematic diagram of the synchronizing circuit. FRAMING OF PICTURE It is not enough that the cylinders on both transmitter and receiver rotate at exactly the same speed; the picture must be Zworykin: Facsimile Picture Transmission 547

framed as well.In other words, the first edge of the picture being sent should be under the spot of light at the transmitter at the same instant that the first edge of the photographic paper is being exposed by the glow tube at the receiver. The framing is accomplished by the following method. The picture to be transmitted is held on the cylinder by a longitudinal

Fig. 12 black band; at the end of the cylinder first transmitted is a narrow white ring. As the light spot explores this ring, a. continuous signal is transmitted except for the interval when the black band is absorbing most of the light. The glow tube at the receiver flashes once for each time the black band occurs, or seven times per second. At the end of the shaft upon which the receiving cylinder rotates is an interrupter which breaks the glow tube current for a time equal to that required for the trans- 548 Dwarykin: Facsimile Picture Transmission mission of the band. If the interruption takes place at the same time the flash normally occurs, the light from the glow tube appears steady, and framing is known to be correct.If the flashes are seen, however, it is necessary to correct the relative position of the glow tube with respect to the position of the cylinder at a given instant. This is done by a process equivalent to rotating the frame of the motor.

Fig. 13 The framing process described above is carried out for each picture transmitted by the intermittent machine, since the framing is lost when the motors are stopped. In the continuous machine the motors run constantly, hence framing is required only at the beginning of transmission. STARTING OF THE RECEIVER When synchronizing and framing is accomplished the picture starts to pass under the transmitter's scanning spot.The Zworykin: Facsimile Picture Transmission 549 starting of the receiving cylinder is accomplished automatically. The principle of operation of this starter is as follows: On the front end of the transmitting cylinder a band of black and white spots is engraved. This can be seen on Fig. 1. When the picture is started, this band comes first under the scanning spot. As a result, the corresponding frequency is produced by the transmitter and reproduced by the receiving amplifier.This frequency operates a small tuned relay, Fig. 11, which in turn starts a grid glow tube.2 The current passing through the grid glow tube operates a lock -in relay which completes the circuit

Fig. 14 to the magnetic clutch; this starts the receiving cylinder. The starter does not require any additional equipment at the transmitter, nor at the broadcasting station. ASSEMBLED MACHINES AND RESULTS OBTAINED Figs. 12 and 13 show the finished appearance of the intermittent type of transmitter and receiver, respectively.Both machines are self-contained, including all the amplifiers, rectifiers, and tuning forks. In size, the cabinets are two ft. square by four ft. high.The transmitter could be installed at any convenient place connected with the broadcasting station by ' See description of grid glow tube by Knowles, Elec. Jour., April, 1928. 550 Zwarykin: Facsimile Picture Transmission means of a telephone line. The receiver should be placed either in a dark room or adjacent to a small developing booth. With the exception of the darkening of the end of the receiving machine for handling the bromide paper, no other precautions are required in the illumination of rooms where both machines are located. In Fig. 14 are shown side by side an original picture and the facsimile transmitted over a short telephone line and a few miles of radio channel. Proceedings of the Institute of Radio Engineers Volume 17, Number 3 March, 1929

NOTES ON GRID -CIRCUIT DETECTION' BY J. R. NELSON (Engineering Dept., E. T. Cunningham, Inc., New York City) Summary-A dynamic method of finding O'i.,18e,,2 or akplaeg, the main term in grid -circuit rectification, is given. This method is based upon formulas given by Chaffee and Browning' and consists in calculating akdoe, from the change of d.c. when a known a.c. input voltage is applied to the grid.The values of akdae, found by the dynamic method are compared with the values found by the usual method. The effect of frequency, internal grid resistance, and external resistance on the detection voltage introduced in the plate circuit are also considered. An experimentally determined curve is given showing the detector frequency distortion of a commercial set for a 2-megohm grid leak and for a i-megohm grid leak. 1HE subject of detection has been covered from the theoretical side in two excellent articles. The article by Chaffee and Browning' gave a comprehensive theoretical treatment of detection together with a bridge method of experimentally determining the detection coefficient. The article by Llewellyn2 covered the general theory of vacuum -tube operation.The theory in the above article was complete and the equations were applied in some detail to grid -leak detectors. An article by Smith' gave a good physical picture of grid rectification as far as the grid circuit itself was concerned. The main difficulty in using the theoretical analysis as given by the first two mentioned articles is in determining the second derivatives of the currents with respect to the voltages over the range of input voltages applied to the detector. To determine these derivatives it is necessary to plot the conductance curve from the static characteristic curve, then plot another curve from the conductance curve.There is also the uncertainty of whether the dynamic value will correspond with the static value found graphically.If the values of the second derivatives Dewey decimal classification: R134. Original manuscript received by the Institute, November 14, 1928. 1 E. L. Chaffee and G. H. Browning, "A Theoretical and Experimental Investigation of Detection for Small Signals," Pxoc. I. R. E., 15, 113; February, 1927. 2 F. B. Llewellyn, "Operation of Thermionic Vacuum Tube Circuits," Bell System Tech. Jour. 5, 433; July, 1926. 3 Lloyd P. Smith, "Theory of Detection in a High Vacuum Thermionic Tube," PROC. I. R. E., 14, 649; October, 1926. 551 552 Nelson: Notes on Grid -Circuit Detection could be found by some simple means for the desired value of input voltage the usefulness of the theory would be increased. The theory itself could be checked and the range over which the theory holds could also be found. Different types of tubes could be compared easily as the main terms in both plate and grid rectification are the second derivatives of the currents with respect to the voltages. It is the purpose of this article to show that the value of the second derivative of the grid current with respect to the grid voltage, the main term in grid rectification, may be found by the change in direct current in the grid circuit with any de -

112 Vo § §"

o -to -0.11 -05 -OA -0.2 0 *02 GRID VOLTAGE Fig. 1-Grid-Current Grid -Voltage Characteristics; Typical C-327 Tube. Eb-45 v. Ef-2.5 v. sired value of input voltage.The results are also a check on the theory. In another article the author' shows how the value of the second derivative of the plate current with respect to the grid voltage, the main term in bias detection, may be found when the input voltage is fairly large. A sensitive direct -current meter with a full scale reading of about five microamperes is required.Good accuracy may J. R. Nelson, "Detection with the Four Element Tube," Piton. I. R. E., 16, 822; June, 1928. Nelson: Notes on Grid-Ctirouit Detection 553 be obtained by balancing out the direct current if it is more than three or four microamperes. There is no difficulty in using a meter of this type in the grid circuit, but it is difficult to keep the plate current constant enough to use it in the plate circuit of a three -element tube. Chaffee and Browning' give the following formulas for the increment of direct current. A2E, 02/= (AE)2 (1) r, 4 ae, Where 46,24= The increment of direct current 02E°= The increment of steady voltage Eg = The maximum value of input voltage r, = Grid resistance Kg= Grid conductance R = Resistance to a steady current 4,2E0= _Th21, (2) r, atc, wig X (rgR) (AK g) 2 (3) 4 ae,

r, 1 X21°_ (AEg)2 (4) r, -FR 4 ae, ax 4A2/, (5) ae, (AE,)2 rg If the d.c. resistance is small compared to r, Eq. (5) reduces to K° L12/, =4 ae. (A2E02 (6) Fig.1 shows the grid -current grid -voltage characteristic of a typical C-327 tube with Eb =45 volts and Ef= 2.5 volts. The curve is steeper than the curves obtained with other types of tubes such as the CX-301 A, CX-340, etc. The tangents are difficult to obtain accurately, but it was felt that if the theory could be checked using this type of tube the method of determining aKdae, by the change in direct current could be applied to other types of tubes without the necessity of checking the values graphically. I Loc. cit., page 129. 554 Nelson: Notes en Grid-Cirouit Detection Fig. 2 shows the grid conductance, 8i,/8e, of the typical C-327 tube, plotted against the grid voltage over the range which is of most interest. This curve was obtained from Fig. 1 by finding the tangent Di,/De, at the different values of grid voltages, then plotting the values of tangents against the grid voltages.Fig. 3A shows 0./C,/0e, or 024,/0e92 plotted against the grid voltage.This curve was determined from Fig. 2 by finding the values of OK,/De, at the different grid voltages.

.....

280

240

200

120

0 -L2 -ID -OA -0.6 -OA -02 0 0.2 GRID VOLTAGE Fig. 2-Grid Conductance, aig/ae, of Typical C-327 Tube. Eb-45 v. Ef-2.5 v. This is the usual method of determining the value of alc,/ se, or 824/0e.2. If the values of 0/c/8eg found from the change of direct current as explained below check the values of ax./ ae, found from the static characteristic as explained above, the assumptions made in developing the theory are justified. Fig. 4 shows the circuit used to determine the values of alCdoe, at different operating voltages.Alternating current from the source flows through the meter A2 and the resistance R. The value of R was less than 10 ohms so that the input voltage is equal to the product of the reading of A2 and R. No load is shown in the plate of the C-327, but if desired the normal load may be placed in the plate circuit.The a. c. source may be either radio frequency or audio frequency, asthe theory is independent of the frequency of DE,.If a specific case is be- Nelson: Notes on Grid-Cirouit Detection 555 ing investigated by placing a load in the plate circuit the input voltage should be radio frequency so as to take into account the input impedance of the detector at radio frequency.

myv

900

600

300 0 /1 SOO

600

1 900 l I;

SI 200 \i SOO -0.96 -OM -064 -0A6 -032 -OAS 0 .04 Ec Fig. 3-Variation of Grid Conductance, alft/ae, of Typical C-327 Tube. Eb-45 v. Ef-2.5 v. The meter Al had a resistance of about 5000 ohms so that it was necessary to use Eq. (5). The value of n, was determined from Fig. 2. The meter Al was shorted for a.c. with a large con -

C..117

Fig. 4 denser so that it was not necessary to correct for AE.After the grid current was above five microamperes it was balanced out.The resistance of the balancing out circuit was high, so 556 Nelson: Notes on Grid -Circuit Detection

that it was notnecessary to correct the increment of d.c.as read on A. The d.c. was balanced out of Al andR was shorted out. The peak volume of a.c.across R was kept constant at 0.035 volt. The short circuiting switch aroundR was then opened, and the increment of d.c.was read on A1.This was repeated for different values of Ec. The values of 824,/ae.2were then calculated from Eq. (5) for the different values of Ec. The valuesthus found are shown plotted in Fig. 3B.The agreement is good considering the difficulty of finding Fig. 3B. The peaksare separated less than the total grid swing, 2X 0.035= 0.07 volt. The dotted portion of B is estimated, asno accurate voltage between - 0.1 and -0.2 of a volt could be determined with theone -volt thermo- voltmeter used. The curves of Fig. 3 do not show wherethe point of maximum detector sensitivity will occur.They show only the value of voltage to which the grid shouldbe biased to obtain the maximum current when the externalgrid circuit is shorted. In determining the point of maximumvoltage on the grid the values of the internal and external gridimpedances as well as the value of alciaei, have to be takeninto account. Chaffee and Browning' give the followingequation for the voltage introduced in the plate circuit.

(Det. E)/= R1 alf . 4[vrtt2cows+ 1 -1-Rilf(2-FRIK.)" aeg (7) agm1BN/mAB 2 +-aep LI)) where (Det. E)1 is the audio -frequencyvoltage introduced in the plate circuit. R1 is the value of the external gridresistor. C1 is the value of the external gridcondenser. Eq. (7) neglects the effect of theinput conductance not due to the electron flow and theinput capacity of the tube at Loc. cit., page 142.Note: If the part due to bias rectification is neglected the same equationmay be obtained from Eq. (43), page 448 of the article by F. B. Llewellyn, "Operationof Thermionic Vacuum Tube Circuits," Bell System Tech. Jour. 5, No. 3;July, 1926. Nelson: Notes on Grid -Circuit Detection 557 audio frequencies.If it is desired to take the above factors into account the article of Ballantine' may be consulted. The voltage introduced in the plate circuit given by (7) may be used as ta, in the usual amplifier equations, and the value of the voltage across the external plate impedance found as soon as its value is known. The usual method of using a heater type of tube such as the C-327 for grid -circuit detection is to return the grid to the

2000

-so -00 -Oh -OA -02 MD VOLTAGE Fig. 5 aK,ag. [ VRI20)2C12-1-1-FR1K,(2-1-141C9) ae, de, CI =250 ppf. A-Dynamic Value of F Zero Frequency B-Static Value of F Zero Frequency C-Dynamic Value of F 3000 cycles cathode which is at zero potential. The grid may be returned to any other part of the B or C supply. It might be of advantage in this type of tube to return the grid to some point of the B supply which is at a small positive potential above the cathode. A larger value of grid leak could then be used to pass through some desirable operating point than if the grid is returned to 7 Stuart Ballantine, "Detection by Grid Rectification," PROC. I. R. E., 16, 593; May, 1928. 558 Nelson: Notes on Grid -Circuit Deteotion the cathode. In this investigation it was decided to return the grid to the cathode and try various values of R1. Fig. 5A and 5B shows the part of (7) in parenthesis plotted for different values of R1 returned to the cathode. Thecurves were plotted by finding from Fig. 1 the required values of R1 to pass through, zero and different values of E,. The value of K, was then found from Fig. 2. A and B were plotted forzero frequency.The presence of the grid condenser causes frequency distortion.Fig. 5C shows curve A when the audio frequency is 3000 cycles and C is 250µµf. There is quite serious frequency distortion if R1 is chosen so that it intersects I, in Fig. 1 at more than 0.7 volt negative. The mutual conductance for this tube is shown in Fig. 6 and the values of ag,/ae, the main term in plate rectification, are shown plotted in Fig. 7A and 7B. Their effect in changing (Det. E): is small over most of the range.The agreement is

INV

1000

SOCI, -12 -W -0.0 -OA -0.1 - 0 et Fig. 6-Mutual Conductance, (Nu of Typical C-327 Tube. Eb-45 v. Ef-2.5 v. not very good between the static and dynamic values.The platecurrent would changeseveralmicroamperes making it difficult to obtain the true change in direct current whena signal was applied to the grid. The dropping off of the mutual conductance as shown by Fig. 6 occurs at a small negative potential.In the usual type of tube the mutual conductance curve straightens out at about 2.5 volts positive, or near the potential of the filament center. The frequency distortion of this tube used in a popular radio set was checked experimentally for two values of grid leaks 2.0 megohms, the grid leak furnished with the set, and 0.5-megohm grid leak. The set uses a push-pull stage for the last audio stage. The frequency distortion was determined experimentally by the method shown by Ballantine.7Briefly this consists of determining the audio amplification from the grid of the detector to the grid of the last tube with the plate Nelson: Notes on Grid -Circuit Detection 559

load or speaker in place.The input is then placed in series with the grid leak with the low side grounded. Because of the shape of the mutual conductance curve it was found to be necessary to modify Ballantine's method slightly.Instead of shorting the grid leak R to determine the audio amplification it was bypassed with a large condenser.This keeps the same bias on the tube for both measurements.This is important if the mutual conductance changes much from the operating

PA liar°Fill1 M rt 1111111E 3. MIN LIM Mil

2 OA eg -VOLTS Fig. 7-Variation of Mutual Conductance, ag,/ae, of Typical C-327 Tube. Eb-45 v. Ef-2.5 v. A-Static Value B-Dynamic Value point to zero bias.After the two curves were determined experimentally the factor to bring them together at a low frequency, in this case 60 cycles, was determined. The values of the curve determined by the input being in series with the grid leak were then multiplied by this factor.The difference between the values at the higher frequency represents the loss caused by frequency distortion of the detector. Curve A of Fig. 8 shows the audio amplification from the grid of the detector up to the grid of one tube of the push-pull 560 Nelson: Notes on arid-Cirouit Detection stage determined as shown above.Fig. 8B shows the curve obtained by placing the input in series with the grid leak after the two curves are brought together at 60 cycles.The grid leak used in this case is the 2.0-megohm grid leak furnished with the set.Fig. 8C shows the same curve after a 4-megohm grid leak was substituted for the 2.0-megohm grid leak. The loss in amplification due to the frequency distortion of the detector is somewhat greater than that shown by C and B of Fig. 5 at 3000

4443

500

400

.....N...... s..."...4.1.:\*"...... s-\\\ \ s 300 r \ 200

a`\- 100 B\

013 LH , 20 30 40 5040.00 SO 200]00400 00000010M 2000 4000 W OO IOD FREQUENCY Fig. 8-Audio Amplification and Detector Frequency Distortion of a Commercial Set. A-Amplification grid of detector to grid second audio tube. B-Detector frequency distortion with 2.0-megohm grid leak. C-Detector frequency distortion with 0.5-megohm grid leak. cycles. This difference is caused by the input capacity of the tube at audio frequencies, which was neglected in calculating B of Fig. 5. The method of determining the values of alC/6e5 at different grid voltages by the change in direct current offers a simple and accurate means of evaluating this factor. The results bear out the theory for small input voltages. No attempt was made to find the range of input voltages over which the theory would hold. The value of grid leak should be chosen to fit the type of tube used and the values of d.c. voltages used on it. A and C of Fig. 5 show graphically the range of grid voltage over which the detector will be most sensitive as far as grid detection is concerned, considering both low and high frequency. For example, according to these curves the detector sensitivity would be about the same from -0.4 to -0.7 volt; above -0.7 volt detector frequency distortion becomes serious. Nelson: Notes on Grid-Cirouit Detection 561

There is another factor to consider also. The conductance at - 0.4 volt is 130X 10-1 mhos and at -0.7 volt itis about 10 X 10-6 mhos. The value of -0.7 volt would be preferable to use as the selectivity of the circuit preceding the detector would be very poor with 130X 10-6 mhos shunted across it.The value of grid leak to use may be found from the static characteristic of Fig. 1 as soon as the operating point is found from Fig. 5. The experimentally determined curves of Fig. 8 bear out the theoretical considerations that the grid leak should be chosen to match the tube characteristics. The grid leak should be as high a value as possible considering only the radio -frequency tuned circuit preceding the detector. The value of the grid leak should be as low as possible considering frequency distortion of the detector. Some value in between will be the best compromise. In the tube analyzed a value of resistance that would bias the tube at about - 0.7 volt would be the best to use. Proceedings of the Institute of Radio Engineers l'olume 17, Number 3 March, 1929

THE RADIATION RESISTANCE OF BEAM ANTENNAS* BY A. A. PISTOLKORS (Radio Laboratory, Ring -Novgorod, U.S.S.R.) Summary-In this paper a new method proposed by Brillouin for the calculation of radiation resistance is applied to several types of beam antennas. New formulas are deduced and some interesting results are obtained showing the distribution of the radiated power among the different wires of beam antennas and giving the numerical value of the radiation resistance in various cases (synphase antenna, antiphase antenna, Marconi three -stage antenna).The radiation resistance in the presence of a perfect conducting plane is also considered. A table of values of the components of radiation resistance is added to the paper for practical use. THERE are two methods of computing the power radiated by an antenna.In the first we calculate the Poynting- vector for each point of space and integrate the normal energy flow through any surface enclosingthe antenna.This method might be called the Poynting-vector method, and has been used in the well known works of G. W. Pierce,' B. van der Po1,2 S. Ballantine,' M. A. Bontsch-Bruewitsch,4 and S. Levin and C. Young.'It may be noticed that we cannot by this method obtain the contributions to the radiated power of different parts of the antenna, which it is sometimes desirable to know when dealing with some practical cases. The other method is based upon the study of the emfs induced in the antenna by the currents in the wires of which the antenna is constructed. Let AB (Fig. 1) be a wire in which flows a current of frequency f and assume that the distributions of the current and of the charges in the wire are known. We may then calculate the electric force in each point M of the space. It depends upon the combined action at this point of allele- ments of the wire. In particular we may take thepoint Mo to lie in the surface of the wire itself and calculate the emf due to the electromagnetic field of the wire.If we assume the action at a distance of the current to be instantaneous,this emf would * Dewey decimal classification: R125.6. Original manuscriptreceived by the Institute, May 25, 1928. Revised manuscript received, October 8,1928. 1 G. W. Pierce, Proc. Amer. Acad. 52, 192, 1916. 2 Balth. van der Pol, Jr., Proc. Phys. Soc. of London, 13,217, 1917. 1 Stuart Ballantine, PROC. I. R. E., 12, 823, 1924; 15,245, 1927. 4 M. A. Bontsch-Bruewitsch, Annalen der Physik, 81,425, 1926. 6 S. A. Levin and C. J. Young, PROC. I. R. E., 14, 675,1926. 562 Pistolkors: Resistance of Beam Antennas 563 be a purely reactive one and we may speak of itas arising from the capacity or the inductance of the wire.If the wire is of a length comparable with that of the electromagneticwave we must take into account the propagation velocity of the field. Then at a moment t there will act in the point M thecurrent and the charge which have existed at the element dxr at the earlier r moment I-7, r,being the distance of the point from dxi) and c the light -velocity. Similarly for the element dx2we must use the values of current and charge existing at the earlier r2 moment(t --c . Under these conditions the calculated emf (or, strictly speaking, the corresponding potential drop) will have a watt component which may be called the radiation emf.

Fig. 1 The product of this emf and the current inan element of the wire gives the radiation from the element. Wemay find by integration the expenditure of power for radiation in the whole antenna or in its different parts. We shall call this method the induced emf method.It was proposed by Brillouine and applied by Kliatzkin" in the analysis of the radiation of a vertical earthed wire.It is based upon the electromagneticfieldequations in the form employing the retarded potentials of Lorentz.8 This paper deals with the radiation resistance of antennas, composed of parallel half -wave vibrators. Radioelectricite, April, 1922. Telegrafia i telefonia bez provodov (TITBP) 1(40), 33, 1927. 8 Lorentz, "The Theory of Electrons," 2 ed., Chap. 1, paragraphs 13 and 14. 561 Pistolkors: Resistance of Beam Antennas

1. Outline of the Method. Radiation EMF. We have first to solve the following problem: a single wire with a known distribution of current being specified, a formula is to be found for the component of electric force parallel to the wire at any point of space. This may be derived from the expression given by Lorentz: aA E= grad.0 ; (1) c at where 1 1 =- [p]dv (2) 47r r is a scalar potential at the given point due to the fixed charges distributed in the space and

1 1 A =- - [pv ]dv (3) 4irc r is a vector potential due to the charges, moving with the velocity v. Applying this to the case of a very thin straight wire and passing from the Lorentz units to the absolute system of units we obtain 1 j" Icrl 4)=- dx (4) E 0

I ili-r/ A= /AC dx (5) where a is the charge per unit length, i is the current in the element dx of the wire, / is the length of the wire, and r is the distance of the given point from dx. The values of charge and current must be taken at the time(t-

Let the origin of coordinates lie at the one end of the wire and let the OX axis lie along the wire (Fig. 2). The component of the electric force parallel to the wire will then be at the point M (a, E) as follows: Pistolkors: Resistance of Beam Antennas 563

1 aAdo 1 IA ai(s_r,4 a 1 go_ric) E = -- --- = - dx dx (6) c at dt r at aE 0 Er If the angular frequency is w then i = I= sin cot and agt-r/c) -coli cosco(t--r (7) dt c ai(t - r/c) 1ar. . r ace -44= dt - cosco(t- - (8) ax w ax c auai from the equation --=- giving the relationship betweenthe atax

Fig. 2 current and the charge along the wire. The instantaneous value of electric force will be as follows cos (wt-mr) ea= -pc[m2 I .dx Mr a cos (cot- mr) al z (9) at o mr axJ w 2w where m= -= - and r = d2 (x - C X Assume that the current is distributed sinusoidally along the wire and is zero at the origin of coordinates. Thus 566 Pislolkors: Resistance of Beam Antennas = /0 sin mx (10) where /0 is the amplitude of the current at the loop. The expressionfor ed will then be ftcos (cot - mr) ea =pc[nt2 /0sin mxdx 0 mr ar cos ((a - mr) I dx] (11) +m-cosat mr mx After integration we obtain: (wt - mrl_E) cos (wt - ea=Acio[cos cos ml (12) 7.1_1 re

where ri_k = \,/d2-1-- 02 and ri .\/=. d2 t2.

Fig. 3 In the particular case in which the point M lie's on the wire itself, d =0 and we shall have: cos (cut- ml+mt) cos (cot- eo= µdo cos ml (13) /-E J. 2. Case of Two Parallel Wires. We shall now study the problem of the radiation of power from a system formed by two parallel wires. For this purpose we may assume some conditions which simplify the solution. We shall consider (Fig. 3) equal Pistolkors: Resistance of Beam Antennas 567

wires whose lengths are multiples of the half -wavelength.The wires are not displaced in height, that is, their ends must lie on the straight line perpendicular to the direction of the wires. The distribution, the phase and the values of current we assume to be identical in both wires. Under these conditions the tangential component of E at any point along either wire comprises two parts, the first produced in each one by its own current and a second produced by the current in the other wire: i.e., e = eo-F ed.This electric force produces an emf in the wire, and the power needed to suppress it will be the radiation power. For the element dx of the wire this power will be dPz = - EL cos orkdx (14) where E and Ix are the effective values of the electric force and current and ¢ is the phase angle. The total power for one wire having the length 1 will be

PE = -fEL cos chdx =-fEo/x cos chodx o (15)

fEdix cosOddX = P 0+ P d - 0 We shall first calculate Pd. Let the full expression of E and I= be written. To obtain the power in watts we must take the current in amperes and Ed,E0 in volts. Then /x = /0 sin mx (16) where /0 is the effective value of the current at the loop in amperes. Ed is also the rms value and from the formula (12) we obtain two components of it

cos ml 1 E'30/0 and E" = 30/0- (17) ri_x r x Each of these components has a different phase angle chi and 412. Let us find them. From the expression (12) we have

ed' = Ed1 cos (Ca- Mr I- El sin(cut- mrz-x-F-2) (18) 568 Pistolkors: Resistance of BeamAntennas

The current in the wire is i= = I, sin cot

Therefore cki = mri_,--and 2

71- cos cki = cos( mri,- = sin mri, (19) 2 Similarly cos cfio = sin mrz. Thus we obtain for Pd ' /sin mrt_x sin Pd=-30102 cos ml sin mx dx (20) f0 r, By integrating we obtain the following expression for the radiated power when the length of the wire is a multiple of the half -wavelength. Pa=30102[2Ci md-Ci m(Vd2 +12 +1) -Ci m(V d2 +12 -1)]=30102Ma (21) Here Ci(x) denotes the integral cosine, d is the distance between the wires, l is the length of the wire. Pa is merely one of the components of radiation power, depending upon the current in the other wire.The second component Po we may obtain as limit of Pd when the distance between the wires approaches 0. Po=liml Pal =30102(E +log 2m1-Ci2m1)=30/02M0(22)

where E =0.577 is the Eulers constant. The whole radiation power in one wire will be P=Po+Po (23) and the radiation power of the system of two wires Pz =2P =2P0+2Pa (24) Dividing the expressions (21) and (22) by ./.2 gives the so- called "radiation resistance." Obviously we may speak of this quantity only when the currents in both wires are equal. 3. Application to Beam Antennas. Synphase System. We shall now apply these results to the computation of the radiation resistance for some types of beam antennas. We shall consider Pistolkors: Resistance of Beam Antennas 569 first the so-called synphase antenna, composed of single parallel vibrators (Fig. 4).The vibrators are situated along a straight line at a distance of a half -wavelength from each other.Their currents are equal and in phase.In this case the beam has a direction perpendicular to the line of the wires. The power radiated by any individual antenna wire is composed of the power due to its own current and that due to the electric force induced in it by other vibrators. We shall denote

Fig. 4 by Pd the power corresponding to induction by a vibrator at a distance d; in the case of the antenna considered d will be a multiple of a half -wavelength. As stated above in (21), Pd=30/o2Md As the currents in all wires are equal the component of radiation resistance'due to another wire will be Rd=303 d The values of these components are given in Table A (line 1) for distances d which are multiples of the half-wavelength.9 Using this table let us now compute the radiation resistance for the antenna composed of three vibrators. For each of the extreme wires we shall have: RI= R3= Ro+RX12±RX= 73.3-12.4+4.1=65.0 II and for the middle wire: R2= Ro+21?),/2=73.3-2X12.4=48.5 fl. The radiation resistance of the whole antenna will be: R3=3R0+ 4RAI2+2RX= 178.50. For these calculations I employed the integral -function curves especially plotted by Mr. E. D. Milovidoff of the staff of the Nijni- Novgorod Radiolaboratory.Interpolation from commonly used tables (Jahnke u. Emde) is rather misleading. 570 Ptstolkors: Resistance of Beam Antennas

For the four -wire antennawe obtain by analogous calculations: R2 = 4R0+6Rx/2+4RA+2R3x/2 and generally for an antenna composed ofn wires: Rn = nR0-1-2(n- 1)Rxii-F2(n- 2)R2),12+ -1-2R(,1)A/2 (25) Table I contains values of the radiationresistance (a) for each vibrator, (b) for the whole antenna,and (c) the mean value for one vibrator. It might be noticed thatwhen increasing the number of vibrators the last quantity isvery rapidly approaching the limit (near 56 ohms), whichwas obtained for the case of an infinitely great number of vibrators byusing the "Poynting vector method" of radiation resistance calculation."

TABLE I Values of radiation resistance in ohms for synphase beamantenna. n -number of wires: R. -resistance of nth wire; R =total resistance. R..=average resistance per wire. n Rg R, RI IN R, AN ler R Rot 2 60.9 60.9 ------121.8 60.9 3 65.0 48.5 65.0 ------178.5 59.5 4 63.2 52.6 52.6 63.2 ------231.8 58.0 5 64.4 50.9 56.7 50.9 64.4 -- -- 287.3 6 63.8 52.0 57.4 55.0 55.0 52.0 63.6 341.4 7 64.0 51.2 56.9 56.1 53.2 56.1 51.2 64.0 395.6 56.5

4. Continuation. Antiphase Beam Antenna. Weshall now pass on to the antiphase antenna (Fig. 5) which radiatesa beam directed in the plane of antenna.It differs from the synphase

Fig. 5 antenna only in the fact that the currents in the adjacent wires have a phase difference of 180 deg.All the radiation resistance components having d equal to an odd number times the half- wavelength must therefore be multiplied by -1.The other components are the same as before because vibrators spaced by an integral number of wavelengthsare in phase. 14 M. A. Bontsch-Bruewitsch. Lc., p. 434. Pistolkors: Resistance of Beam Antennas 571

The radiation resistance for this antenna is expressed by the following general formula:

= nR 0 - 2( n- 1)RA/2+ 2( n- 2)RA - + /2 (26) where the quantitiesRdmay be taken from Table A (line 1). The results of calculations for the antiphase antenna are given in Table II. From Tables I and II we may see that the radiation resistance is different for the various wires in the antenna and this difference is unequal for the two types of antennas. As the number of Vibrators is increased the difference diminishes.

TABLE II Value of radiation resistanoe in ohms for antiphase beam antenna (for explanation see legend of Table I).

n Rs A6 AN Rs Rs Rs R, R E.

2 85.7 85.7 ------171.4 85.7 3 89.8 98.1 89.8 ------277.6 82.5 4 91.5 102.2 102.2 91.5 ------387.2 96.8 5 92.7 103.9 106.2 103.9 92.7 -- -- 499.2 99.8 6 93.4 105.1 108.0 108.0 105.1 93.4 -- 612.8 102.1 7 93.9 105.8 109.1 109.7 109.1 105.8 93.9 727.2 103.9 5. Parallel Wires Displaced in Height. As a next step in the development of this method, the radiation resistance of parallel vibrators displaced in height may be calculated, (Fig. 6).The investigation of this case will enable us to study the radiation resistance in the presence of a perfectly conducting plane and to calculate the radiation resistance of multistage antennas.

Fig. 6 We will deduce a formula for the radiation power due to the emf induced in the vibrator B by another vibrator A. The wire B is placed a distance d and elevated on a height h with respect to A. We shall denote this power by P(d, h) and the corresponding radiation resistance component by R(d, h). The electric force near B due to the first vibrator is defined by (12). 572 Pie°Mors: Resistance of Beam Antennas

The law of current distribution in the wire B isas follows: /.= /0 sin m(x- h) (27) Proceeding exactly as in Sect. 2 we shall obtain for the radiation power due to the induction the following expression: "4/2 (sin mri_ssin mr P(d, h) =30102 sin m(x-h)dx. J. n_. (28) The integration gives for the corresponding radiation resistance component a rather complicated expression, as follows:

X R(d, h)= -15 sin mh[S(d,h--)-2S(d,h) 2 (29) -FS(d,h-F-)]

X - 15 cosmh[C(d,h--)-2C(d,h) 2

+C1 d,h-F 2 where S(x, y) and C(x, y) are the functions:

S(x, y) = Si 112(112T2:FY2+Y) - na(N/ Y 2 -11) (30) C(x, y)=Ci m(N/x2+ y2+ y)+Ci m(N/7-1-e-y).

Fig. 7 This formula is the most general one for thecase of two parallel vibrators. The expression (21), obtained for the vibrators placed at the same height, is a particular case of it for h =0. For the -other particular case, when d= 0 (Fig. 7) we find: Pistolkors: Resistance of Beam Antennas 573

X R(0, h)=-15 sin mh[Si 2m(h- 2Si 2mh 2

+Si 2m(h-F-X)] 2 (31) h2 -15 cos mh[log Ci2m(h --X ) h2-x2 2 4

-2Ci2mh+Ci 2m(h+-X 2 which is in agreement with the analogous formula obtained by M. A. Bontsch-Bruewitsch. Having any given complex antenna system comprised of n synphase parallel vibrators we can by means of expressions (29) calculate the radiation resistance for every one of them. This resistance will be:

Rz = R(0, 0)+R(d1 110+ R(d2, h2)+ +R(dn_i, h_1) (32) where h and d denote the height difference and the distance between the first vibrator and each other one; R(0,0) = -d A)

Fig. S

6. Antenna Over Perfectly Conducting Plane. The expression (29) may also be used for the calculation of the radiation resistance of an antenna erected over a perfectly conducting plane by application of the simple image theory. We shall treat the ease of an antenna of which the vibrators are placed at the same height over the plane and at equal distances d from each other. 574 Pistolkors: Resistance of Beam Antennas

Let ho be the height of antenna over the conducting plane (Fig. 8).Introducing the correction due to the images we shall obtain: For each of the extreme wires: RzA=73.3+R(0,h)+R(d,0)+R(d,h)+R(2d,O)

+R(2d,h)+ +R[(n-1)d,0]+R[(n-1)d,h] (33) X

where h=2110+-i

For each wire second from the edge RzB = 73.3 + R(0,h)+2 [R(d,0)-1- R(d,h)]

+ -FR[(n-2)d,0]-1-R[(n-2)d,h]. (34) If the radiation resistance of the whole antenna is to be found we can use a formula analogous to the formula (25)

Rz = nRo-F2(n-1)R1-1-2(n-2)R2-1- -F2R,,--1 (35) where RkR(--kX 0)+R(? -X 2 2 2 The calculations were carried through by the author for a

X X 3X X aynphase 7 -wire antenna elevated 0,- --, and - over 8 4 8 2 the plane. The results are shown in Table III. We may notice that with increasing height the total radiation resistance rapidly

TABLE III

he v., ter, Ti...ri flmrs T4 R r,,,

0 84.7 58.2 74.7 82.0 497.3 71.0 X 65.0 41.8 59.8 43.7 376.8 53.8 13 X 82.8 44.7 58.8 42.2 i 374.8 53.5 3X i 85.8 50.5 59.0 51.2 401.8 57.4 X 66.4 51.2 55.5 54.4 400.8 57.2 i 84.0 51.2 56.1 53.2 395.6 56.5 approaches the value obtained for free space, but the energy distribution between the individual wires is different. As expected the radiation resistance increases near the plane. Pistolkors: Resistance of Beam Antennas 575

7. Multistage Antenna. The method of induced emfs may be applied to more complicated systems, particularly multistage antennas.Let us consider for example a three -stage antenna of the type employed by Marconi. A unit of such an antenna is a system of three synphase vibrators, spaced along a straight line. These vibrators are connected through antiresonant coils.Let us compute the power due to the emf induced by such an antenna unit in another one spaced at a distance d (Fig. 9). We assume the currents to be equal and in phase.

Fig. 9

The power induced in vibrator I of the wire B may be resolved into three parts due to the vibrators 1-3, respectively, of the wire A. Using our notation we may write:

P Al= P(d,0)+P(d, +(P,X) (36) 12 Similarly for the II and III vibrators:

PAri = P(d,O) +2P(d, -D (37)

P1112--- PAr The whole power induced in the wire B will be 576 Pistolkors: Resistance of Beam Antennas

X PAB= 3P(d, X) +4P(d, 2P (d, X) (38) 2 and generally in the case of n -stage wire

PAB=nP(d, 0) + 2(n- 1)P (d,-0+2(n- 2)P (d, X) (39)

-F2P(d,(n-1)-2-)t ).

262232222215209205 196 173 Fig. 10

The radiation resistance of single three -stage wire is obtained by substituting d=0.

Po =3P(0, 0) +4P(O,2)+2P(0, X) = 317.1 St (40)

249217208204201200192 178

Fig. 11 Pistolkors: Resistance of Beam Antennas 577

In order to calculate the radiation resistance of different wires and of the whole antenna, formulas may be used analogous to those obtained above. The author has performed these calculations for the case of a 16 -wire antenna, the distance between the wires being assumed

to be - The results are shown graphically in Fig. 10; the 2 numbers below give the radiation resistance in ohms. The mean value of this resistance (214 ohms) is marked by a dotted line. Analogous calculations were also made for this antenna elevated X at -over a perfectly conducting earth.The mean value of 4 the radiation resistance for one wire is then 206 Il.The energy distribution is shown in Fig. 11.In both cases this distribution

Fig. 12 is very nonuniform in the extreme wires.This means that the design of the feeding devices for the several wires must be quite different, if we wish to obtain equal currents in all vibrators. 8. Table A.Various other types of antennas may be com- puted in the same manner. To simplify the calculations a table is appended (Table A), containing the functions R(d, h) for values of d and h, which are multiples of a half -wavelength. This table will be found useful in the calculation of many practical types of directive antennas. As an example let us calculate the radiation resistance of an antenna formed by three five -stage wires (Fig. 12) spaced at a 578 Pistolkors: Besistanee of Beam Antennas

distance of a half -wavelength from each other. Proceeding exactly as in Sect. 7 for the three -stage antennawe shall obtain for the radiation resistance component R1 of the wire A dueto the wire B the following expression:

X X =54-x, 0)+ , 6RC--,X\ 2 2 2 2 )

3 +4R -+24-X,2X) (41) 2 2

The values of R(d, h) may be taken from the table.Thus we obtain R1= 5(-12. 36)+8(-11. 80)+6 (-O. 78)+4(+0. 80) +2(-1. 00) = -159.70 (42)

TABLE A

\ d 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 hN 0 +73.29 -12.36 +4.08-1.77+1.18 -0.75+0.42 -0.33 0.5 +26.40 -11.80 +8.83-5.75+3.76-2.79+1.86-1.54 1.0 - 4.065- 0.78+3.56-6.26+6.05-5.67 +4.51 -3.94 1.5 + 1.78 + 0.80-2.92+1.96+0.16-2.40+3.24-3.76 2.0 - 0.96 - 1.00+1.13+0.56-2.55 +2.74 -2.07 +0.74 2.5 + 0.58 + 0.45-0.42-0.96+1.59-0.28-1.59+2.66 3.0 - 0.43 - 0.30+0.13+0.85-0.45-0.10+1.74-1.03

TABLE A (Cont'd)

4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 ...N"si..h 0 +0.21 -0.18+0.15-0.12+0.12 -0.10+0.06 -0.03 0.5 +1.08 -0.85+0.69-0.57 +0.51 -0.45+0.36 -0.30 1.0 +3.08 -2.50+2.10-1.80+1.56-1.18+1.14-1.00 1.5 +3.68 -3.40+3.14-2.90+2.61 -2.31+2.06-1.86 2.0 +0.51 -1.30+1.82-2.24 +2.28-2.29+2.26-2.14 2.5 -2.49 +2.00-1.35+0.49-0.06-0.45+0.85-1.03 3.0 -0.09 +1.12-1.87+1.77-2.02+1.71-1.32+0.66 The values of d and Is are given here in parts of the wavelength; those of radiation resistance in ohms. To find the radiation resistance component R2 due to the wire C or the radiation resistance of the single five -stage wire Ro we must substitute in the above expression d =X or d= 0 instead A d=- Therefore 2 R3=R3=+103.0 0i2 and Ro = 558.5S1 Pistolkors: Resistance of Beam Antennas 579

The radiation resistance of the wires A and C will then be: RA= Rc=Ro-FRi+R2= 501.8 12 and of the wire B

RB=Ro-F2Ri= 239.1 SZ . The whole radiation resistance of the antenna will be 1242.712, the mean value for 1 wire =416.2 CZ and for 1 vibrator =83.0 12. Proceedings of the' Institute of Radio Engineers Volume 17, Number 8 March, 1929

Discussion on SIMPLE INDUCTANCE FORMULAS FOR RADIO COILS* (HAROLD A. WHEELER) R. R. Becher': A formula shown in this paper* for the rapid computation of the inductance of air -core radio coils corresponds to one which the writer has used for about six years. Having found by experience that many engineers to whom this and other simple formulae have been disclosed have an aversion to using them because they have insufficient accuracy, it is believed that further analysis might be of interest. The formula L =an2Q (1) has been called the universal inductance formula since tables, formulas, and charts have been derived to cover almost every shape of air -core inductance in evaluating Q, which is a factor depending entirely on the physical dimensions of the coil.For a single layer solenoid the Lorenz formula may be used or the tables which have been obtained from it for representative values. This formula is one of the most exact in its field. An inspection of a curve for Q plotted from these tables indicates that it is probably of a hyperbolic nature. If so a comparison of the relation between 1/Q and b/2a would be a straight line. Analysis of this line gives an empirical value for Q (after converting to inch measurements by multiplying by 2.54). 100 Q = (2) (b/a +0 . 9) Substituting (2) in (1) gives a1/421000 L = centimeters 10b +9a (3) a1/41 =10b microhenries (4) +9a which is the formula previously reported.The deviation between the curve of 1/Q and a straight line is a measure of the accuracy of this equation. For greater accuracy with very short coils the denominator may be changed to (8.3a +10b). It may be shown that the form of (4) could be obtained from fundamental considerations, and since it brings to light interesting facts it will be derived. One of the basic inductance relations is 4 irtz1 (5) Inductance Reluctance of path For an air core with unit permeability the reluctance is the value of (length of path/area of path). It is this relation that makes the ordinary Proc. I. R. E., 16, 1398; October, 1928. 1 Decatur Mfg. Co., Brooklyn, N. Y. 580 Discussion on Wheeler Paper 581 inductance formula so complicated since both the length and area of the flux path are indeterminate.But the effective length and the effective area values are interrelated so that their quotient is a constant for any one coil.It is convenient, therefore, to assume a value for the area and from this determine the length of the flux path.It may be assumed that the area of the path is the area enclosed by the average turn, i.e., area = ea', throughout the whole length of the flux path. Then the length of the flux path may be taken to be equal to (b +q) where b is the length of the coil and q is a fictitious length equal to the length of the return path of the flux outside of the coil. Substituting these values in (5) gives 4wn2 (2rn)22.54n1 L - 2.54- (8) )b + q b +q 1112 Comparing (6) with (4) indicates that the, flux return path for the values assumed is equal to 0.9a (which may appear paradoxical but the value is due to the small area of the flux path assumed exterior to the coil). This same procedure has been applied to many other types of coils and may be applied to many more. In many cases it is desirable to derive a simple inductance formula for a system of coils which may be under consideration. In one case such a relation was simply obtained for a special shape of coil antennas which assisted materially with the experiments. If a series of measurements are made to cover representative conditions such a formula may be readily derived. (a) Determine the value of 1/Q from (1) using the available data, for several values of 2a/b. (b) Determine (k) from two representative values of Q, say Q' and Q" with their corresponding shape ratios b'/a' and b"/a", using k = (b'/a' - b"/a")Q, (7) (c) Determine q for several values of Q from q =1 /Q -h/ak (8) The value of q should not change very much over a considerable range of values for b/2a. A multilayer coil presents greater difficulties. Several have been disclosed, such as the Brooks -Turner formula, which is probably the one most commonly used. As a first approximation it can be shown that the geometric mean distance from a point within a rectangle to the rectangle is very nearly equal to 0.223 (b +c). Here c is the winding depth. The g.m.d. from a point on a line to the line is equal to 0.223 b. From this it would seem that the factor (b +c) could be substituted in the single layer formula for the factor b, to obtain a multilayer formula: (9) 9a +10 (b -Fc) This formula gives fair accuracy for short coils, but in some other cases a large error results. A more accurate analysis gives: One L (10) 9a +10(b +c)-2bc a 582 Discussion on Wheeler Paper

It has been shown2 that no great error results when a single layer coil inductance formula is used for flat spirals by substituting (c) for (b). Formula (4) may be used. It should be kept in mind that all of these formulas are based on current sheet considerations and are subject to other corrections. They are thus not true "empirical" formulas, which are ordinarily designed to fit a given system of measurements. In conclusion it may be well to call attention to the normal accuracy requirements of an inductance formula.I have seen engineers reject an approximate formula with an admitted error of say 3 per cent and reject slide rule computations in favor of long hand computations, but who never- theless used an ordinary scale graduated in sixteenths to determine the dimensions of the coil.The nominal diameter of the tube was used (3 inches) for 2a, whereas the actual diameter was later found to be 3.045 in., and adding the diameter of the wire increased the value to 3.095 in. The length was taken as 2% in., whereas the actual length (nd). was 2.82 in.The total error due to these measurements was over 6 per cent, although the greatest error in the measurements was less than is in. In this light great accuracy in a formula seems in many cases to be of secondary importance. Harold A. Wheelers: I believe Mr. Hatcher gives insufficient attention to two very important details which I carefully noted for both formulas recommended in my brief paper to which he refers.First, what is the maximum error of the approximate formula? and secondly, within what limits of shape factors does this maximum error obtain?Until these questions are answered, no engineer is justified in using an approximate formula. The type of formula under discussion has no theoretical basis for "short" coils (b+c <2a), since only a small part of the total flux links with all the turns. Furthermore, when there are few turns or large spacing between turns, which is often true of "short" coils, one or two corrections cannot be neglected. Theoretical consideration of a coil of very small winding space yields the following formula of a different type: an2 .9a1 L microhenries (4) 13. 5logob +e-4 (dimensions in inches). On the assumption of uniform current distribution in a rectangular cross-section, the error of this formula is less than 3 per cent when b+c5a. Formula (4) is the only type which remains accurate as the term b-Fc decreases indefinitely. The Hazeltine formula (1) is the most generally accurate of any of its type, as applied to multilayer coils of medium and small winding space. Since preparing the above paper, I find that in July, 1920, I first derived formula (2) therein, with almost identically the same constants now recommended. The Hazeltine formula (1) therein was derivedabout 1916 and published shortly thereafter. 2 J. A. Fleming, "Wireless Telegraphist's Pocketbook," p. 97. s Hazeltine Service Corporation Laboratory, New York City. Proceedings of the Institute of Radio Engineers Volume 17, Number 3 March, `9 9

MONTHLY LIST OF REFERENCES TO CURRENT RADIO LITERATURE THIS is a monthly list of references prepared by the radio laboratory of the Bureau of Standards and is intended to cover the more important papers of interest to professional radio engineers which have recently appeared in periodicals, books, etc.The number at the left of each reference classifies the reference by subject, in accordance with the scheme presented in "A Decimal Classification of Radio Subjects-An Extension of the Dewey System," Bureau of Standards Circular No. 138, a copy of which may be obtained for 10 cents from the Super- intendent of Documents, Government Printing Office, Washing- ton, D. C. The various articles listed below are not obtainable from the Bureau of Standards. The various periodicaLs can be consulted at large public libraries. R100. RADIO PRINCIPLES R113 Radio transmission and the upper atmosphere (editorial). Experimental Wireless and Wireless Engineer (London),5, 657-659; December, 1928. (Abstract of a paper by Appleton comparing the three methods for determining the effective height of the Heaviside layer. The frequency change method, the angle of incidence method, and the group retardation method should give the same equivalent height of the ionized layer.) R113 Baumler, M. Feldstarkemessungen auf grosse Entfernungen im Rundfunkwellenbereich. (Field strength measurements at great distancesin the broadcasting range). Elektrtsche- Nachrichten Technik, 5, 473-477; November, 1928. (Report of cooperative field intensity measurements at Konigsberg, Hamburg, Karlsruhe, and Dresden. All these stations took readings of waves (190, 405, and 585 meters) arriving in the form of long dashes (30 seconds duration) from a Fend- ing station located at the Reichspostamt at Doeberitz. Day and night effects are reported.) R113 Sreenivasan, K. Vber die Wellenausbreitung in einem disper- gierenden Medium. (On the wave propagation in a dispersed medium). Zeits. far Hochfreguenztechnik, 32, 121-124; October, 1928. (It is shown that the group velocity of high -frequency waves varies for different frequencies when passing through a dispersed medium such as the Heaviside layer.It seems, therefore, to be evident that a modulated wave on account of its two side bands should produce distortion after passing through this ionized layer.) R113 Fuchs, J. Der Einfluss der Erdatmosphare auf die Ausbreitung kurzer Wellen. (On the influence of earth atmosphere on the propagation of short waves). Zeits. fur Hochfrequenztechnik, 32, 125-129; October, 1928. (It is shown that the strength of the received signal for short waves after passing over sea water depends on the distribution of the pressure of the atmosphere. From this it follows that the atmosphere produces scattering similar to disused reflection.) 583 584 References to Current Radio Literature

R113.5 Goldstein, S.The influence of the earth's magnetic field on electric transmission in the upper atmosphere.Proc. Royal Soc. (London), 121A, 260-285; November. 1928. (Based on lectures by Prof. J. Larmor. The theory of the effect of the magnetic field of the earth on the propagation of electromagnetic waves in the Heaviside layer is given in much detail.) R114 Schindelhauer, F.Ober elektromagnetische Storungen.(On electromagnetic disturbances). Elektrische-Nachrichten Technik, 5, 442-449; November, 1928. (Study of the clicks and grinders by means of the direction finder due to Watson - Watt. The author concludes that since the direction of the maximal disturbance is either along or perpendicular to the earth's magnetic axis, most of the atmos- pherics are due to field changes above the surface of the earth. These field changes cause the electron to be drawn from the sun towards the earth and then produce the eddies of the Heaviside layer. The first causes the clicks and the latter the grinders.) R114 Watson -Watt, R. A.Present status of knowledge of atmos- pherics. Experimental Wireless and Wireless Engineer (London), 5, 629-652; November, 1928. (Reviews work done on this subject by himself and others up to present date.) R120 Moser, W.Die Obertragung der Energie vom Sender zur Antenne bei kurzen Wellen.(The transfer of energy of short waves from the transmitting set to the antenna).Elektrische- Nachrichten Technik, 5, 422-426; November, 1928. (Description of the system carrying the high -frequency power to various individual antennas used for beam transmission. The parallel wire and the concentric tube system is used for feeding the power into the antennas and a method is described by means of which the losses of the distributors can be found.) R125.6 Meissner, A. and Rother, H. Ober die Bestimmung des gtin- stigsten Ausstrahlwinkels bei horizontalen Antennen. (On the determination of the favorable radiation angle in horizontal antennas).Zeits. f. Hochfrequenztechnik, 32, 113-115; October, 1928. (The most favorable radiation angle for 15 and 20 meter wavelengths was determined for horizontal polarisation at the center using horizontal multiple antennas in connection with a parabolic reflector.It was found that the most favorable radiation happened when it took place along the tangent of the surface of the earth.) R125.6 Gresky, G.Die Wirkungsweise von Reflektoren bei kurzen elektrischen Wellen.(The operation of reflectors for short electric waves).Zeits. fur Hochfrequenztechnik, 32, 149-162; November, 1928. (The beam effect of a vertical antenna for the case of a cylindrical parabolic reflector and a plane reflector (several vertical wires) along a straight wire is experimentally studied.For the parabolic reflector the ratio of focal length to wavelength should be 0.27 and for the plane reflector 0.2. The tuned reflectors give smaller dimensions.) R125.6 BOhm, 0.Die Bundelung der Energie kurzer Wellen.(The concentration of the energy of short waves). Elektrische- Nachrichten Technik, 5, 413-421; November, 1928. (Explains the beam transmission system employed by the Telefunken Co. A very clear presentation of the underlying principles giving at first the radiation characteristic of the dipole, then that of a group of dipoles along a straight line and in a plane.) References to Current Radio Literature 585

R125.6 Gothe, A.ttber Drahtreflektoren. (On wire reflectors). Elektrische-Nachrichten Technik, 5, pp. 427-430; November, 1928. (Description of the action of reflector antennae.Wire reflectors reduce the strength of the backward beam considerably. Complete screening by mesns of reflectors is only possible if the radiation coupling between antenna and reflector is variable so that the amplitude and the phase of the reflector current can be properly adjusted.) R130 Podliasky.Equilibres instables et regimes statiques parasites dans les circuits electriques associes aux triodes.(Unstable equilibrium and regular static parasites in electric circuits associated with tubes). L'Onde Electrique, 7, 475-487; November, 1928. (Conclusion of the paper on pp.287-306 of the July, 1928 issue of this periodical.) R131 Rajski, C. Les capacities internee de la lampe a plusieurs electrodes.(Interelectrode capacities of multi -electrode tubes). L'Onde Electrique, 7, 461-474; November, 1928. (Expressions are derived for the interelectrode capacities of electron tubes taking the space charge into consideration.If the tube is not burning the usual inteielectrode capacities (filament -grid, filament -plate, and grid -plate) are observed but when the filament is emitting electrons it is necessary to consider four capacities, the grid capacity, the plate capacity, the grid -plate capacity, and the plate -grid capacity.) R134.75Boella, M. Sur le calcul des amplificateurs a moyenne frequence superheterodyne. (On the calculation of the intermediate frequency amplifier stages of a superheterodyne).L'Onde Elec- trique, 7, 500-508; November, 1928. (Analytical treatment of the amplifier stages of a superheterodyne used for the amplification of the intermediate frequency.) R134.75Ramsay, J.F. A double superheterodyne.Experimental Wireless and Wireless Engineer (London), 5, 669-672; Decem- ber, 1928. (Description of a two -fold superheterodyne. The first superheterodyne changes the received high frequency to a 600-kc current which is then amplified by two stages of radio -frequency amplification after which another heterodyne produces a 150-ko current. Tnis is passed through three stages of radio -frequency amplification rectified and amplified by a two stage audio -frequency amplifier.) R144 Waite, G. R., Brickwedde, F. G., Hall, E. L.Electrical resistance and magnetic permeability of iron wire at radio frequencies. Physical Review, 32, 967-73; December, 1928. (Discussion of the results of B. Wwednensky and K. Theodortschik and those of the authors of this paper who could not detect a critical variation in the resistance of iron wire in the vicinity of 3000 kc.) R144 Jackson, W. The effect of frequency on the value of high resistances of the grid leak type.Experimental Wireless and Wireless Engineer (London), 5, 677-679; December, 1928. (The very high resistance of a grid leak consists in reality of a pure resistance with a small capacity (a few aka) in parallel. It is shown that above 10' cycles/ sec. the effective resistance changes and the parasitic capacity current becomes pronounced.) R200. RADIO MEASUREMENTS AND STANDARDIZATION R250 Moullin, E. B.An ampere meter for measuring alternating currents of very high frequency.Proc. Royal Soc. (London), 121A, 41-71; November, 1928. 586 References to Current Radio Literature

(Gives the theory and construction of a new high -frequency ammeter which is based on the repulsion between two parallel conductors carrying the current to be measured. The frequency effect can be calculated. One conductor is fixed and the other one can move against a small elastic constant. This motion is a measure of the repulsion force and therefore of the current.It is noted by means of a microscope.) R261 Aiken, C. B. A sensitive vacuum tube voltmeter. .Jnl. Optical Soc. of American and Review of Sci. Instruments, 17, pp. 440- 450; December, 1928. (A vacuum -tube voltmeter is described which utilises the heterodyne principle for obtaining increased sensitivity.) R300. RADIO APPARATUS AND E 1IIIPMENT R342.15Koehler, G.The design of transformers for audio -frequency amplifiers with preassigned characteristics. PROC. I. R. E., 16, 1742-1770; December, 1928. (Requirements of ideal transformer are stated and difficulties encountered in attempting to build transformers for interstage coupling units which will meet these requirements are pointed out.) R343 Kiipfmiiller, K.'Ober die Dynamik der selbsttlitigen Ver- starkungsregler.(On the dynamics of the automatic amplifier stabilizers). Elektrische-NachrichtenTechnik,5,459-467; November, 1928. (A system is described by means of which the amplified intensity is automatically kept constant.Based on the principle developed a receiving set has been built which produces the same output intensities during times at which the input voltage (due to fading) varies up and down.) R344 Eller, K. B. On the variation of generated frequency of a triode oscillator due to changes in filament, current, grid voltage, plate voltage, or external resistance. PROC. I. R. E., 16, 1706-1728; December, 1928. (General expressions developed for generated frequency of grid -tuned and plate - tuned generators.) R344.4 Ritz, M. Essais sur ondes tres courtes.(Tests on very short waves). L'Onde Electrique, 7, 488-499; November, 1928. (Study of transmission of waves of 2 to 8 meter length. Gives generator diagrams. The experimental results agree with those due to R. Mesny.) R344.4Uber eine Methode zur Erzeugung von sehr kurzen elektro- magnetischen Wellen.(On a method for the production of short electromagnetic waves).Zeits. fur Hochfrequenztechnik, 32, 172; November, 1928. (Description of magnetron oscillator for the production of very short waves, X=29 cm.) R359 Hahnemann, W. Uber die neuere Entwicklung des Maschinen- senders fur kleine Wellenlangen. (On the new development of machine transmitters for short wavelengths). Elektrische- Nachrichten Technik, 5, 431-437; November, 1928. (Description of the latest development of toe Lorentz alternators with frequency multiplication.The improvements consist in producirg frequencies in the broadcast band; filters for reducing the effect of the side bands; increase of the life of the frequency multipliers and reduction of the Thriller effect which causes a periodic change in the frequency.) R376.3 Wolff, I. Sound measurements and loudspeaker characteristics. PROC. I. R. E., 16, 1729-41; December, 1928. (Description given of methods used to measure loud speaker response. Typica characteristic curves given.) References to Current Radio Literature 587

R400. RADIO COMMUNICATION SYSTEMS R412 Bailey, A., Dean, S. W., Wintringham, W. T. The receiving system for long wave transatlantic radio telephony. PROC. I. R. E., 16, 1645-1705; December, 1928. (Determinations show that frequencies near 60 ko are beat suited for transatlantic radio telephone transmission.Various types of antennas described. Mathematical discussions of wave antenna, antenna arrays, and probability of simultaneous occurrence of telegraph interference are givin in lipendices.) R500. APPLICATIONS OF RADIO R526.1 Stowell, E. Z. Unidirectional radio beacon for aircraft. Bureau of Standards Journal of Research, December, 1928.Research Paper No. 35. Reprint copies obtainable for 10 cents from the Superintendent of Documents, Government Printing Office, Washington, D. C. (Description of one of the schemes tried out by the Bureau of Standards for unidirectional radiobeacon work.Directive and non -directive fields are transmitted simultaneously with the proper phase and amplitude relations in order to obtain unidirectional effects.) R570 Birnbaum, H. W. Die Fernlenkversuche der Reichsmarine in den Jahren 1916-1918. (The guiding experiments of the German Marine in the years 1916 to 1918).Zeits. fur Hoch- frequenztechnik, 32, 162-170; November, 1928. (A description of the system used for guiding airplanes by meats of radio.) R592 Crawley, C. A year's progress in commercial wireless. Wireless World and Radio Review, 23, 801-804: December 12, 1928. (Discusses automatic S.O.S., position finding,. eae jamming of broadcast, beam telephony.) R800. NON -RADIO SUBJECTS 534 Ballantine, S. Note on the effect of reflection by the microphone in sound measurements.PROC. I. R. E., 16, pp. 1639-1644; December, 1928. Physical Review, 32, 988-992; December, 1928. (Attention is called to the fact that ordinary microphones will not indicate the true pressure of an uadieturbed sound wave for the entire audio -frequency band. The correction can, however, he found by employing a standard spherical mounting of which the diagram occupies a small area at the pole. A method of this type can, therefore, be used instead of obtaining the calibration curve with the Raleigh disk.) 534.3 Watanabe, Y. fiber die vermittels einer Stimmgabel erregten Rohrenoszillatoren. (On tuning fork vacuum tube oscillators). Zeit& fur Hochfrequenztechnik, 32, 116-121; October, 1928. (The equations for these oscillators are derived and the mechanical as well as electrical oscillations are compared in order to give an expression for the frequency obtained in terms of the true frequency of the tuning fork.) 535.3 Barnard. G. P.The selenium cell: Its properties and applications.Jnl. I. E. E. (London), 67, 97-120; December, 1928. (Gives the historical review on the work done with the selenium cell and describes the several factors affecting the conductance.Gives applications to photometric and relay problems and shows applications to the optophone, photo - phone, talking film and television. An extended list of references is given at the end of this paper.) 537.55 Pardue, L. A., Webb, J. S.Ionic oscillations in the glow discharge. Physical Review, 32, 946-49; December, 1928. (A detailed experimental study of ionic oscillations in the glow discharge which was originally found by Widdington and Appleton.) 588 References to Current Radio Literature

537.65 Tawill, E. P.Nouveau mode de developpement d'electricite par torsion dans lea cristaux de quartz.(New method of production of electricity by torsion on quartz crystals). Comptes Rendus, 187, 1042-1044; December 3, 1928. (A way was found of producing charges on a quartz cylinder axis along the optical axis when applying a torsion about this axis. Suggests calling it atrepho- electricity. because it is different from ordinary piezo-electricity.Shows that for a twist in one direction charges of opposite polarity appear on the surface of the envelope of the cyclinder and the faces perpendicular to the axis.A twist in the opposite direction reverses the polarity. The polarity also depends on the optical rotation.) 537.65 Mandell, W. The determination of the piezo-electric moduli of ammonium Seignette salt.Proc. Royal Society (London), 121A, 130-140; November, 1928. (Theory and determination of the pieso-eleotric constants of ammonium Seig- nette salt.) 621.374.2 Landon, V. D. A bridge circuit for measuring the inductance of coils while passing direct current. PROC. 1. R. E., 16, 1771- 1775; December, 1928. (Bridge circuit described in wbieh inductance of oo;I is compared to resistances and a capacitance.) Proceedings of the Institute of Radio Engineers rolume 17, Number 3 March, 1.929

CONTRIBUTORS TO THIS ISSUE Beers, G. L.: Born 1899 at Indiana, Pennsylvania. Received B. S. degree, Gettysburg College, 1921. Westinghouse graduate student course and engineering school, 1921-22. Radio engineering department, Westing- house Electric and Manufacturing Company, specializing on superheterodyne development, 1922 to date.Associate member, Institute of Radio Engineers, 1927. Carlson, Wendell L.: Born 1897 at Jamestown, New York. Grad- uated Bliss Electrical School, 1918. Associated in radio receiver development for the Navy as a civilian at Washington Navy Yard, 1918-23. Associated with broadcast receiver development for the General Electric Company, 1923 to date.Received Charles A. Coffin award, General Electric Company, in1925 for accomplishments in superheterodyne development. Associate member, Institute of Radio Engineers, 1919. Clapp, James K.: Born December 30, 1897 at Denver, Colorado. With Marconi Wireless Telegraph Company, 1914-16; U. S. Navy, 1917-19, foreign service, 1918-19; Radio Corporation of America, 1920, also 1922-23.Received B.S. degree, Massachusetts Institute of Tech- nology, 1923; instructor in electrical communications, M.I.T., 1923-28; M.S. degree, 1926.Engineering department, General Radio Company, Cambridge, Mass., 1928 to date.Associate member, Institute of Radio Engineers, 1924; Member, 1928. Harris, Sylvan: Born May 5, 1898 at Philadelphia, Pennsylvania. Received E. E. degree, University of Pennsylvania, 1922; engaged in power maintenance work, Midvale Steel Ordnance Company, Phila- delphia, 1917-18; Leeds and Northrup, 1918; technical editor, Lefax, 1922-23; managing editor, Radio News, 1923-25; director of research and design, Stewart -Warner Speedometer Corporation, Chicago, 1926-27; laboratories of Brandes Products Corporation, Newark, N. J., 1927 to date.Associate member, Institute of Radio Engineers, 1920; Member, 1926. Hector, L. G.: Born December 15, 1894 at Clarendon, Pennsylvania. Received A. B. degree, Oberlin College, 1920; M. A., department of physics, Columbia University, 1922; Ph. D., Columbia University, 1924. Private research assistant to A. P. Wills, 1920-22; Tyndall Fellow, 1923- 24.Instructor in physics, Oberlin College, 1922-23; assistant professor, 1924-27; professor of physics, University of Buffalo, 1927 to date. Author of "Modern Radio Receiving," and of papers on magnetic measurements of gases. Member Sigma Xi, American Physical Society, and American Association for the Advancement of Science. Associate member, Institute of Radio Engineers, 1926. Vice-chairman of the Buffalo -Niagara section. Hulburt, E. 0.: Born October 12, 1890. Received Ph. D. degree in physics, Johns Hopkins University, 1915.Taught undergraduate and graduate courses at Western Reserve University, Johns Hopkins Univer- sity, and University of Iowa. With A. E. F. in France as lieutenant and later captain of signal corps.At present superintendent of Heat and 589

590 Contributors to this Issue

Light Division, Naval Research Laboratory, Bellevue, D. C. Author of a number of papers on experimental and theoretical work in spectroscopy, physical optics, and radiotelegraphy. Kozanowski, H. N.: Born August 15, 1907 at Buffalo, New York. Received B. S. degree, University of Buffalo, 1927.Graduate student, department of physics, University of Buffalo, with teaching assistantship, 1927-28. At present graduate student, department of physics, University of Michigan. Maxis, H. B.: Born 1885.Received A. B. degree, University of Michigan, 1909; M. S. degree, 1910; Ph. D. degree, Johns Hopkins University, 1927. Associate professor of physics, Birmingham -Southern College, 1922; professor of physics, Emory and Henry College, 1923; consulting physicist, Naval Research Laboratory, 1925 to date, working on researches, photo -elastic studies, and theoretical studyof the upper atmosphere. Nelson, J. R.: Born October 27, 1899 at Murray, Utah.Power inspector, Western Electric Company, 1922-23; B.S. in E.E., University of Southern California, 1925; Engineering Record Office, Bureau of Power and Light, Los Angeles, Calif., 1925; radio test, Radio Development Laboratory and Tube Research Laboratory, General Electric Company, 1925-27. M.S. degree in E.E., Union College, 1927. Engineering Depart- ment, E. T. Cunningham, Inc., July, 1927 to September, 1928. Research Laboratories, National Carbon Co., September, 1928 to date.Associate member, Institute of Radio Engineers, 1927. Pistolkors, A. A.: Born October 11, 1896 at Moscow, Russia. Grad- uated from First St. Petersburg Gymnasium, 1914. During World War served in Radio Corps, Caucasian Army. Chief of Bakou Radio Station, Caucasus, 1919-20; chief and instructor, Post Office Radio School at Vladikavkaz, North Caucasus, 1920-23. Entered Moscow High Techni- cal School, 1923; received E. E. degree, 1927. From 1926 to date with Lenin Radiolaboratory, Nizhni-Novgorod. Smith -Rose, R. L.: Born at London, England. Studied physics and electrical engineering, Imperial College of Science and Technology, 1912- 15; graduated, B. Sc. degree, 1914.Received Ph. D. degree, London University, 1923; D. Sc. degree, London University, 1926.Telephone engineer, Messrs. Siemens Brothers and Company, Limited, 1914-19. Appointed to scientificstaff of the National Physical Laboratory, Teddington, England, 1919; placed in charge of wireless division of the laboratory in 1923. Past lecturer in various technical institutions. Member of several committees of the Radio Research Board, England. Zworykin, Vladimir K.: Born 1899 in Russia. E. E. degree, Petrograd Institute of Technology. In 1912 entered the laboratory of the College de France in Paris, where he worked on research in X-rays under Professor P. Langevin. During World War served in Russian Army as an officer on the Radio Corps. In January, 1919 came to U. S. A., and since 1920 has been a member of the Research Laboratory of the Westinghouse Electric and Manufacturing Company. Received Ph. D. degree, University of Pittsburgh, 1926; title of thesis, "A Study of Photo -Electric Cells and Their Improvement." The Finest Receivers Are Thordarson Equipped

"4' "" "Thordarson products have been chosen for incor. poration in Federal Ortho-Sonic Radio Sets because we have always been certain that we would receive a quality of product entirely in keeping with the high standard set by us for Federal Receivers."

President, Federal Radio Corporation THORDARSON ELECTRIC MANUFACTURING CO. Transformer Specialists Since 1895

Huron, Kingsbury and Larrabee Streets Chicago

BORDARSONFLA DI 0 TFLAN SFORMERS SUPREME IN MUSIC _t I. 1:41 It MANCE csl bo When writing to advertisers mention of the PROCEEDINGS will be mutually helpful. The Super Model Pacent Phonovox

Type 106B: New Super Phonovox with balanced,/ tone arm.Bronze finish. 0.5.00 Type 106A:New Super Phonovox without tone,/ arm.Bronze finish. 012.00 Type106C:Ultra Phonovox combination. Heav425.00 gold finish. Again Pacent presents-a full season ahead, as usual-an astonishing advance in pick-up design-the electrical radio phonograph reproducing instrument that will be acceptedas standard. 1. No rubber bearings-assures 4.Mechanical filter-no peaks, freedom from wear, eliminates minimum surface noise. variationandmaintainsex- 5. Tone box :tinges backwards, quisite tone quality. facilitates needle insertion. 2.New balanced tone arm assures perfect contact with the 6. Low centerofgravity-no record-no scratching. resonance or rattle. 3. Higher frequencyrange- coversentirerangeoffre- quenciesfoundonphono- graph records. PACENT ELECTRIC CO., INC. 91 SEVENTH AVENUE, NEW YORK CITY Pioneers in Radio and Electric Reprod.iction for over20years Manufacturing Licensee for Great Britain and Ireland: Igranic Electric Co., Ltd., Bedford, England

When writingto advertisers mention of thePROCEEDINGSwill be mutually helpful. SM For the New Tubes around which will centet 1930 quality receiver designs Silver -Marshall Now Offers The 720AC Screen -Grid Six and 669 Power Unit

Designed to utilize to the fullest extent the superior characteristics of the new RCA and Cunningham A.C. heater -type screen -grid tubes as radio frequency amplifiers, and the equally advantageous high output of the new intermediate power tube without the use of high plate Complete 720AC KIT, without cabinet, tubes, or power voltages, the 720AC offers at once unit, $78.50. Type 700 metal shielding cabinet, $9.25. an advanced view of 1930 recep- WIRED complete in cabinet, tion. $108.00. New Power Transformersand Chokes Power Transformer (B--1 besides two extra mitre of tapstoworkintoany normal S -M 335 power transformer,as seriesorseries -parallelarrange- used in the 669 power unit for ment of one or more standard new A. C. screen -grid and inter- magneticordynamicspeakers. mediate power tubes, has one 5 - Will handle over 20 watts with- volt2 ampererectifierfilament out core saturation. winding, two 21/4 -21/2 -volt 6 ampere filament windings, and two 248, open -mounted..... $7 high voltage secondaries. With one 228, in shielding case $8 '80 tube,it delivers 300 -volts at 100 m.a.Two mounting styles: Unichoke 335, cast iron end mountings $15 Atwo -winding high -inductance 335U, open bracket mounting $15 filter choke, tuned by the Clough resonanceprincipletoeliminate Universal Output Choke 120cycleripple,withonly7 A choke designed especiallyfor mfd.capacity. Inductances 2 couplingtwopowertubesin henries and 30 henries,at110 push-pull ('71A, '50, or the new m.a.load. Continuousrating intermediate tube). Provided with 125 m.a. two end leads and a center tap 331 Unichoke $8

Ask for the new Silver -Marshall General Catalog, containing full descrip- tionofallthe above new products, the new Clough -system push-pull transformers and many other recently added items.

LosaicuAl.G3ACK. SO.NBLVD. SILVER -MARSHALL, Inc., U.S.A.

When writing to advertisers mention of the PROCEEDINGS will be mutuallyhelpful. Nit "RS :IIRaytheonE

IDIADGIALif IN TELEVISION Raytheon Laboratories took up the task of developing tubes for television apparatus as soon as practicable principles of television transmission and reception had been worked out. Raytheon progress in this field has, therefore, been concurrent with television development as a whole. Foto Cell We now offer, as equipment of proved Television sending tube efficiency, the Foto-Cell sending tube and in bard vacuum or gas filled types the Kino-Lamp receiving tube. The Foto-Cell has been developed to the point where cells are made which will respond to various frequencies in the light -spectrum. The Kino-Lamp is being produced in numerous types and styles, which provide suitable light -sources and light-sensitive relays for all systems. These include various types of spot -glow lamps, as well as flat -plate type-all of which will glow in white, blue, green and various tints of orange. These developments are also effective in phono-film, in sound -reproduction, and in all systems where a sensitive light - relay or a sensitive light -source is needed. We invite correspondence and welcome opportunities to extend co-operation in television a and allied developments Ki11914n1P Television receiving tube RAYTHEON MFG. CO. adapted to all systems CAMBRIDGE, MASS. Price $7.50 When writing to advertisers mention of thePROCEEDINGSwit! be mutually helpful. Iv PRECISION FILAMENT RIBBON and WIRE (uncoated) for Radio Vacuum Tubes

Thousands of Feet of Accuracy V MINIMUM SHRINKAGE V CORRECTLY SPOOLED (unwinds freely) V UNIFORM SIZE, WEIGHT, & RESISTANCE V ACCURATE (within .00005" of specifications) SIGMUND COHN, 44 GOLD ST., NEW YORK

When writing to advertisers mention of thePROCEEDINGS willbe mutually helpful. V REVOLUTIONARY

When writing to advertisers mention of the PROCEEDINGS will be mutually helpful. VI RADIO PRODUCTS

Designed especially for Radio Manufacturers THE ISOLANTITE COMPANY OFAMERICA announces the production of a new line of completely assembled radio products.Guided by a desire to do full justice ro the superior qualitiesofIsolantiteasan insulating substance, three groups of products, featured herewith, have been evolved which are outstanding in many respects.This plan of centralized manufacture and assembly makesitpossible toofferIsolantitequality insulationincompletedarticles and revolutionary designsathighlyattractiveprices. 1-THE NEW ISOLANTITE SOCKETS Revolutionary in design, construction and method of assembly in the set chassis, these Isolantite sockets mark an epoch in the development of radio set assembly. They combine the advantages of sockets now in use with many radical improvements, both electrical and mechanical. Their unusually great physical and dielectric strength, their low inter- element capacity, low leakage and the permanence of the contact prongs are in themselves features which make Isolantite sockets outstanding. Coupled with these advantages is the entirely new and novel method of keyed mounting which makes for the utmost in speed and economy of set production and security of the assembly. 2-THE NEW ISOLANTITE NEUTRALIZING CONDENSER Isolantite comes to the fore with advanced principles of neutralizing condenser construction in offering this unique essential.Moreover, because Isolantite will not warp, shrink or otherwise change in shape or size, the new Isolantite condensers fixed for indefinite periods.Its rock -like structure provides a permanent base which is fabricated to receive punched plates "formed" into place on the mcunt- ing. No nuts, screws or rivets are required in its construction.These condensers provide maximum insulation, low leakage and minimum volume losses. 3-THE NEW ISOLANTITE ADJUSTING TOOL A suitable tool which may be used as a socket wrench as well as a screwdriver is offered by Isolantite.This tool is Isolantite insulated, is available in various lengths and will prove highly practical in the factory, service station and laboratory. Wrench ends are available for use with standard 8-32 hexagonal nuts as well as 6-32 nuts.Both the wrench and screwdriver are securely fixed to the Isolantiteshaft which is hexagonal in shape. The attention of Radio Set and Accessory Manufacturers is invited to these new products.They are described in detail in Isolantite Bulletin No. 100C. A copy will be sent upon request.

71gyicalARInsidation [solantice Company ofAmerica (Incorporated) New York Sales Offices, 551 Fifth Avenue When writing to advertisers mention of the PROCEEDINGS will be mutuallyhelpful. VII "ESCO" Synchronous Motors for Television

In addition to building reliable and satisfactory motor generators, "Esco" has had many years of experience in building electric motors for a great variety of applications.

Synchronousmotors, small, compact,reliable,selfstarting are now offered forTelevisionequip- ment.They require no direct current for excitation, are quiet running and fully guaranteed. Other types of motors suitable for Television may also be supplied.

Write us about your requirements.

Machines for operating 60 -cycle A.C.RadioReceivers,Loud Speakers and Phonographs from Direct Current Lighting Sockets Without Objectionable Noises of any Kind. The dynamotor. and motor generators are suitable for radio receiver, and for combination instruments containing phonographs and receivers.Filters are usually required.The dynamotor' and motor generator, withfiltersgiveas good or better results than are obtained from ordinary 60 -cycle lighting sockets. They arefurnished completely assembled and connected and are very easily installed. These machines are furnished with wool - packedbearingswhichrequirevery little attention, and are very quiet running. Dynamotor with Filter far Radio Receivers Write for Bulletin No. 243-C.

How can "ESCO" Serve You? 11110ELECTRIC SPECIALTY COMPANY TRADE "ESCO" MARK 300 South Street Stamford, Conn.

When writing to advertisers mention of the PROCEEDINGS will be mutually helpful. VIII pOWERIZER Power and super Power i11IIMPLIWILMS

Our many years of We have in tone reproduction ex- production a very mod- perience is at your erate priced disposal. line of amplifiers using the new UX- 245 Tube. As a matter of STANDARD UNITS economy, we PX-2 2 Stage -226 will try to meet and 210. PXP-210 2 Stage - your inividual 226 and Push -Pull requirements 210 Second Stage. PXP-250a 3 Stage with our PXP-171amplifier. ard amplifiers. of about watts undistorted output with more than usually balanced output However, we over the audible tone will be more range. than pleased Send us your special problems and we will send you Bulletin to take up 1026-I, together with our your special recommendations. problems. Licensed by Radio Corporation of America and Associated Companies RADIO RECEPTOR COMPANY, Inc. 106 SEVENTH AVENUE New York, N. Y. 307 N. Michigan Blvd. Chicago, Ill.

When writing to advertisers mention of thePROCEEDINGS willbe mutually helpful. IX Surpasses even itself !

Peerless Insulation today is a product of evolutionEvery property possessed by the first sheet of Peerless has been developed and intensified by scores of improvements during the forty years this remarkable insulation has served industrial America. The fabric as produced today has a much higher tensile strength.It easily folds with or across the grain without cracking at the fold. Absolute uniformity in thickness has been achieved, as well as a fine, even finish. A new cnemical tees went has made it denser and harder, too, without im pairog the high dielectric strength which has always char. actcized Peerless Insulation. Check up on the service requirements of your procucts. If any of them need the advantages offered by Peerless, use Peerless; for only in this super -paper will you find all these desirable properties present in their relative imp's -Lance. NATIONAL VULCANIZED FIBRE CO. Wilmington, Del., U. S. A. Offices in principal cities PEERLESS INSULATION

When writing to advertisers mention of the PROCEEDINGS will be mutually helpful. X That is what radio executives, pur- chasing agents, and engineers said when they read this startling announcement... Polymet smashes the neck of the bottle

THE coil situation of 1928! Remember it, or don't you like to? It was the wrench In the spokes. the neck of the bottle for 1928 radio and speaker production. And now Poly met. the same Polymet long famous for Polymet Condenser. and Resistant... smash.. the neck with crud. which will be heard throughout coll.uing industries. POLVMET MAKES COILS! The high Quality, quiek Service, and absolute Dependability. long 'roamed with Polymet Condensers and Resiet aaaaa are now carried into the roil industry. The Conlon Electric Manta. factoring Company of Easton, P.., toil -makers for over eleven years, has been acquired. From this date iti. Polymet plant, under Polymet management, making PolyCoils. to Polymet apecifteations. Polymet is ready to. and can. end your roil POLY COILS problems, wh aa they may be. Blue print. of mmanufacturers' requirements are especially rolirited and will receive immediate attention. Audio T re aaaaa inert, Polymet Manufacturing Corp. Pwe rt aaaaa me". 591Broadway, New York City nyolooases, 1,141 Collo for POLYMET PRODUCTS

When writing to advertisers mention of the PuocEEmuGs will be mutually MN:. XI Radio Potter Products Electrical Condensersa.of Merit

Sound Investments Always Pay Dividends QUALITY in the Potter Condenser is given by the use of highest grade paper, foil and impregnating wax. LONG LIFE has been attained by manufacture in a factory devoted to the ex- clusiveproductionofcondensersby special processes. UNIFORMITY is essential to give best results in any radio receiver or power amplifier and is given by careful and skilled workers. A series of tests during the making and rigid inspection controls the production. ECONOMY does not always come with the purchase at the lowest price.The additionalcostisthe investment that pays dividends by reducing the repair charges which are sure to grow if condensers fail under operating conditions.

Potter Condensers include a full line of By -Pass, Filter and Filter Block Condensersforalloftherequired capacities and working voltages. Special attention is given to manufacturers arranging condensers to meet their requirements.Recommendations and quotations will be gladly made covering your condenser problems. A Condenser Assembly For Every Use The Potter Co. North Chicago, Illinois 4 National Organisation at rw Bergin

W hen to advertisers mention of thePROCEEDINGSwill be mutually helpful. XII New613ookc-----= Modern Radio Reception By CHARLES R. LEUTZ

384 Pages -250 Illustrations -6x9 Fully Bound

SOME OF THE SUBJECTS COVERED Radio Laboratory Apparatus Super Heterodynes A Current Supplies Short Wave Reception Vacuum Tubes Power Packs Radio Definitions A/C Tubes Resistance in Radio Power Amplification Radio Frequency Amplification Long Wave Reception Radio Measurements Trouble Finding Oscillators Shield Grid Tubes

An ideal book for the radio experimenter, engineer, service man or anyone interested in broadcast reception. Price $3.00 Postpaid Unless you are entirely satisfied, the book can be returned within five days after receipt and your money will be refunded immediately. C. R. LEUTZ, Inc. LONG ISLAND CITY, NEW YORK, U.S.A.

When writing to advertisersmentionof the PROCEEDINGS will be mutually helpful. XIII HELPTIOPUBLIC DECIDE TO BUY YOUR PRODUCT NE thing is certain in the highly competitive Radio Omarket.The public is confronted with a multiplicity of claims concerning the performance ability of the many Radio sets offered to them. The day of experimenting is 'passed; the pinnacle of perfection is near.The set of tomorrow which will gain the radio public's favor is the one that has the inherent quality and manufacturing superiority to lift it a little higher than the others. Scovill has pioneered in the production of Radio Con- densers. Constant laboratory research backed by modern facilities for the making and testing of radio condensers has helped it to reach a state of manufacturing perfection that appeals to many of the leading radio manufacturers. Today itis aiding manufacturers to build that better set that the public demands.It can help you. Every step in the manufacture of Scovill radio partsisunderstrictlaboratorysupervision. OVILL MANUFACTURING COMPANY WATERBURY, CONNECTICUT NEW YORK BOSTON CHICAGO PROVIDENCE PHILADELPHIA CLEVELAND Los ANGELES SAN FRANCISCO CINCINNATI ATLANTA DETROIT In Europe,-THE HAGUE, HOLLAND

ll'hcn writing to advertisers mention of thePROCEEDINGSwill be mutually helpful. XIV The Pyrohm is Built to Carry the Load!

AROVOX Fixed and Tapped Vitreous Enamelled Pyrohm Resistorsare made in a wide range of resistance values and wattage ratings to suit everypower supply requirement. They are built to thesame high standards as Aerovox Mica Condensers, SocketPower Con- densers and Filter Condenser Blocks.

The August issue of the AerovoxRe- search Worker containsan interesting and instructive articleon How to Calcu- late Voltage Dividers for PowerSupply Devices.A copy will be sent freeon request.

"Built Better" 78 Washington St., Brooklyn, N. Y.

When writing to advertisers mention of the PROCEEDINGS will be mutually lietpfI ul. XV INSTITUTE SUPPLIES EMBLEMS Three stylesofInstitute emblems, appropriately coloredto indicate the various grades of membership intheInstitute, trA are available. The approximate size of each emblem isthat of the illustrations. The lapel button is of 14k gold, the background being enamelled inthe membership color,thelettering being gold. The buttonissupplied with a screw back with jaws which fastenitsecurelyto the coat.This style emblem canbeobtainedfor$2.75,postpaid,forany grade. The pinisalso of 14k gold.Itisprovided with a safety catch and is appropriately colored for the various grades of membership.Price, for any grade, $3.00 postpaid. The watch charm is handsomely finished on both sides and is of 14k gold.The charm is equipped with a suspension ring for attaching to a watch fob or chain.Price for any grade, $5.00 postpaid. BINDERS Thebinderpictured herecontainsover three inches of filing space.It is designed to accommodate a year's supply of PROCEEDINGS.It serves either as a temporary transfer binder or as a permanent cover. Itis made of handsome Spanish Grain Fabrikoid in blue and gold. The binderisso constructedthat

each individual copy of the PROCEEDINGS willlieflatwhen the pages are turned.Copies can be removed from the binder in a few seconds and can be permanently preserved in undamaged condition.Hundreds of these binders are sold each year.Price, $1.50 each, or $2.00 with the member's name or the PROCEEDINGS Volume Number stamped in gold. BACK ISSUES OF THE PROCEEDINGS Back issues of the Proceedings are available in unbound form for the years 1918, 1920, 1921, 1922, 1923, and 1926 at $6.75 per year (six issues). Single copies for any of the years listed to 1928 are $1.13 each.For 1928 (where available) the single copy price is $0.75.Foreign postage on the volume is $0.60 additional.On single copies $0.10. A number of individual copies for years other than those listed above are available.For a list of these, members should apply to the Secretary. Bound volumesinBlueBuckram bindingareavailablefortheyears 1920, 1921, 1922, 1923, 1925, and 1926 at $8.75 per year.The bound volume for 1928 is priced at $9.50.Foreign postage is $1.00 per volume. Bound volumes, for the above years, in Morocco Leather binding are avail able at $11.00 each. These prices are to members of the Institute. FOURTEEN YEAR INDEX The PROCEEDINGS Index for the years1909-1926, inclusive,isavailableto members at $1.00 per copy.This index is extensively cross indexed. YEAR BOOK A copy ofthe current year book will be mailed to each member, when available.The 1927 and 1928 year books are available to members at $0.75 per copy, per year. When ordering anyoftheabove,sendremittancewithorderto The Secretary, The Institute of Radio Engineers, 33 West 39th Street, New York, N.Y.

When writing to advertisers mention of the PROCEEDINGS will be mutually helpful. XVI Karas Precision Parts

For Set Manufacturers

SINCE 1894 the. Karas Electric Co. has been identified with the manufacture of apparatus of the most exacting precision, electrically and mechanically.As early as 1900 we were making radio parts for experimental purposes, and since the advent of broadcasting in1922, millions of Karas precision parts have been used in buildinit quality radio sets.Karas variable condensers and audio transformers have a reputation unequalled in the trade. Manufacturers who use Karas parts in their sets are assured instantly of a nation-wide acceptance for high quality.The better grades of radio setswillprofit by this association. Karas Variable Condensers Now Available to Manufacturers The famous Karas variable condenser is now available to manufacturers of radio sets on a volume production basis, including3- and 4 -gang jobs,retaining the efficient characteristics synonymous with the Karas name. Karas Audio Transformers Manufacturers models of Karas audio transformers, known in' radio circleslottheir precision construction and dependability, are also available to set manufacturers in any designs desired. Manufacturers, send us your specifications for Variable Condensers and Audio Trans- formers to meet your complete production requirements or as an additional source of supply.Our central location and large manufacturing facilities assure you of exceptional service. KARAS ELECTRIC COMPANY 4040-60 N. Rockwell Street, CHICAGO, ILL.

When writing to advertisers mention of the PROCEEDINGS mill 6e mutusIty helpful. XVII Fixed and Adjustable Resistors for all Radio Circuits

BRADLETUNITa RADIO manufacturers, set builders and experimenters demand reliable resistors for grid leaks and plate coupling resistors. For such applications Bradleyunit-B has demonstrated its superiority under all tests, because: 1-.Resistance values are constant +-Adequate current capacity irrespective of voltage drop across 5-Rugged,solid - molded con- resistors.Distort ion is thusavoided struction 1-Absolutely noiseless 3-No aging after long use e-Easily soldered. Use the Bradleyunit-Bin your radio circuits

RADIOSTAT This remarkable graphite compression rheostat. and other types of Allen-Bradley graphite disc rheostats provide stepleas, velvet - smooth control for transmitters, scanning disc motors and other apparatus requiring a variable resistance.

LABORATORY RHEOSTAT Type E-2910-for general laboratory service. Capacity 200 watts. Maximum current 40 amperes. A handy rheostat for any laboratory. Write for Bulletins Today! Allen-Bradley Company 282 Greenfield Avenue Milwaukee, Wisconsin

When writing to advertiscrs mention of thePROCEEDINGSwill be mutually helpful. XVIII VALLEY PARTS Punchings and Castings Pressings and Deep Drawings Plating and Spraying Cotton Interweave Coils Permanent Magnets Set and Speaker Parts Sub -Assemblies of all descriptions B -Cast Chokes A -Metal Laminations THE modern manufacturer of radio equipment requires a source of supply that is equipped- in machine and experience-to manufacture parts and accessories of intricate and exacting design. To be of most value the fabricator of parts must know what lies back of the blue print of the parts or accessories. We supply that need. We have served and are continuing to serve the largest radio manufacturers in the country, and we are ready to quote you from your blue prints on the smallest laminations or a complete magnetic circuit for your Dynamic Speaker. Our Engineering Department is ready to cooperate with you in the solution of your problems.

Send in your blue prints. VALLEY APPLIANCES, Inc. 634 Lexington Ave. ROCHESTER, N.Y.

Members R.M.A.

When writing to advertisers mention of thePROCEEDINGS willbe mutual*, helpful. XIX has just published new Bulletin No. K -81o, covering a very complete line of radio fre- 3W' quency ammeters from too M.A. to I00 amperes. There are three different sizes, 372", 4" and 772". Send for your copy of new Bulletin No. K -81o.

"Over thirty years' experi- 4" ence is back of ROLLER -SMITH"

ROLLER -SMITH CO., 2134 Woolworth Bldg. NEW YORK, N.Y.

WORKS: Bethlehem, Penna.

Offices in Principal Cities in U.S.A. and Canada.Representatives in Australia, Cuba and Japan. 734"

When writing to adverti.ters mention of thePROCEEDINGSwill be mutually helpful... XX The HAMMARLUND "Battleship" Condenser Sections Matched to Within 1/4 of One Percent

Sections can be balanced with absolute precision, with Ham- marlund Equalizing Con densers directly attached to recesses provided in the "Battleship" Chassis. CREAL one -dial control is assured by using the new Ham- marlund "Battleship" Multiple Condenser. Two, three and four -gang models, with accurately matched sections, provide the finest possible capacity tuning for modern radio receivers.Slight variations in individual circuits can be perfectly compensated for by directly attaching Hammarlund Equalizing Condensers to each unit. Rugged "Battleship" frame; oversize shaft, except on dial ends and every feature of the famous Hammarlund "Midline" Con- denser make this the prince of multiples. Made in two capacities -350 mmfd. and 500 mmfd.-2, 3 and 4 gangs.Ask us to quote on your requirements. Add Hammarlund Prestige to Your Own HAMMARLUND MANUFACTURING CO. 424-428 W. 33rd St., NEW YORK

.0,1-'89..ttet.Racticy

PRECIS/ON aammarlundPRODUCTS

When writing to advertisers mention of the PROCEEDINGS wilttie mutually helpful. XXI NewQuick Economical !

Three Assembly Operations and this Job is Done!

Tighten two nuts, make one soldered connection and the new Eby Combination Antenna and Grid strip is completely assembled. No insulating washers-no lining up holes. Ground post automatically grounded-Antenna post automatically insulated. Furnished with soldering lug and nut assembled on Antenna post.

Samples and quotations on request

The H. H. EBY MFG. CO., Inc. 4710 Stenton Avenue, Philadelphia, Pa. Makers of Eby Binding Posts

Irhen writing to advertisers mention of the PROCEEDINGS will be mutually helpful. XXII OIL IT SEEMS THAT OUR CON- DENSERS WHICH ARE IM- PREGNATED DIRECTLY IN OIL ARE BEING CONFUSED WITH OTHER SO CALLED "OIL" PROCESSES. THE LIFE OF OUR OIL IM- PREGNATED CONDENSER IS SO SUPERIOR AS TO BE ABOVE COMPARISON WITH OTHER TYPES AND IN JUSTIFICATION OF OUR PRODUCT WE FEEL THIS ANNOUNCEMENT IS NEC- ESSARY. PCRICICON

CONDENSER CORPORATION OF AMERICA 259-271 CORNELISON AVE. JERSEY CITY, N. J. When writingto advertisers mention of thePROCEEDINGS will bemutually helpful. XXIII AComplete and frompthfulatinfAcerrice

FORMICA provides promptly uniform and high qualityphenol laminatedinsulatingsheets, tubes and rods, and parts of any type that itis possible to machine from them. The equipment both for producing the fundamental shapes and for machining them toyour blue prints is the largest and most complete in the industry. The Formica plant is centrally located where the promptest delivery is possible to the largest number of factories. Send your drawings for quotations THE FORMICA INSULATION CO. 4626 Spring Grove Avenue, Cincinnati, Ohio oRMICA Made hom Anhydrous Bakelite Resins SHEETS TUBES RODS munumnimmummounimmommurnmmuumenimnommonunmummitionimmiimmumminummiumniiiiimminimiin When writing to advertisersmention of the PR., i, if' I" If f11,:l',1,1,,11! XXIV Details! Your radio designs this season will be judged by details-refinementsim- provements, adjustments, accurate balance, and better results.And that is precisely where accurate resistance control comes into play. With the CLAROSTAT line, you can provide the necessary finished performance for your radio sets.Forinstance: Why not a real volume con - Volume Control trot, capable of anything from a whisper to the full output of the set?The Volume Control Clarostat does just that- and itis inexpensive, compact, neat, readily mounted and foolproof. Yes, there should be atoneas well Tone Control as a volume control.The live -wire set and loud -speaker manJfacturer is going to have a tone control-the equivalent of the pedals on the usual piano.The Volume Control Clarostat can be used as a tone control with certain condenser combinations we'll tell you about. Plate voltages and grid bias particularly with Voltage Control voltages,screen -grid tubes, must be accurately adjusted.The DUPLEX CLAROSTAT attends to that end.Its screwdriver adjustment prevents tampering by lay hands.

rThe background noises Hum -Controlmustbereducedto an absolute minimum. Center -tap transformers won't do.Crude potentiometers won't do.But there's the Hum- Dinger to solve the problem.Neat.Compact. Inexpensive.Screwdriver adjustment. for details on how to improve your radio designs.And if you are in Willie charge of designor production on radio sets and loudspeakers, write on your firm letterhead and we shall place you on our mailing list for technical bulletins, samples and so on from time to time.Also, don't hesitate to send along your special resistance problems. Clarostat Manufacturing Company, Inc. R .11Y1 A Specialists in Radio Aids 285-7 North Sixth St. Brooklyn, N. Y.

When writing to advertisers ment;on of thePROCFEDI NGSwill be mutually helpful. XXV 0

@Announced the A. C. Screen Grid Tube, Type AC -22 the screen grid tube usingthe separate heater principle and requiring 1.75 amps at 2.5volts.

CeCo pioneering is done without the fanfare of trumpets but itis pleasing to know that many engineers look with confidence to the laboratories of this organizationfor each new development in the tube industry. . . .a reward not measured in profits.

Do not miss CeCo'sentertaining radio broadcast each Monday evening at 8:30 Eastern time (7:30 Central time)over the Columbia Broadcasting System.

CeCo Mfg. Co., Inc., Providence, R.I. CoC)Radio

I ;ling to advertisers mention of the PaoceFoiNcs will be ,itutually XXVI-rubes WHOLESALE RADIO HEADQUARTERS

Abreast of the New Developments in Radio No industry in the world's history has at- tracted so many inventors and experimenters as the radio industry.Something new is always on tap.Contrast the old wireless days with the modern electrically operated talkingradio. Think of whatisstillto come when perfected television, telephony, short wave control, etc., are fully realized. In keeping with the policies ofWholesale Radio Headquarters (IV.C. Braun Com pany), our service lies in testing out and determining which ofthesenewest marvels are practical, salable and usable for the greatest number.Our task isto study the multitude of new merchandise, select those items that are thoroughly proved and reliable, and make it easy for the public NEW LINES to secure these while they are still new. SPRING and SUMMER A huge and varied line of standard radio merchandise is carried in stock for quick shipment to all RADIO SETS, KITS, PARTS parts of the country.This service assuresthe SHORT WAVE, TELEVI- dealer and set builder ofeverything he needs,all SION, SPEAKERS, obtainable from one house, without shopping around SUPPLIES, PORTABLE at dozens of different sources.It saves considerable RADIOS and time, trouble and money.For example, when you PHONOGRAPHS want a complete radio set or parts for a circuit, AUTO TIRES and you also will want a cabinet, loud speaker, tubes and other supplies and accessories.You know that ACCESSORIES at Braun's you can get everything complete in one order, and thus save days and weeks of valuable ELECTRICAL GOODS time, besides a considerable saving in money. Wiring Fixtures, Esc. New Lines for Spring and Summer SPORTING GOODS Here, all under one roof,is carried the world's Outing Clothing, Baseball, largest stocks of radio sets, kits, parts, furniture, Golf Goods, Etc. andsandaccessori e s foithefor seasonrradpeport- an able radioss phonographs summerrado a complete line of auto tires, tubes and supplies, HOUSEHOLD electrical and wiring material, camping and outing SPECIALTIES equipment, tents, golf goods, sporting goods;in Vacuum Cleaners, Phono- fact, a complete merchandise line to keep business graphs, Electrical Toys humming every day, every week and every month in the year. Do You Get Our Catalog? If you don't receive our catalog, by all means send us a request on your letterhead to insure getting each new edition as promptly as it comes out.Braur.'s Big Buyers' Guide is crammed full of bargains and money -making opportunities that you cannot afford to pass up. W. C. BRAUN COMPANY Pioneers in Radio 600 W. Randolph St., Chicago, Illinois

When writing to advertisers mention of the PROCEEDINGS wll be mutually helpful. XXVII C-6hetrend is toward ALUMINUM

The latest Grigsby. Grunow condenser Alcoa Radio Sheet (at right). I.rig.by- seas first developed Grunowhas always forAtwater Kent. used Alcoa Radio The latest Atwater Sheet for its vari- Kent condenser as- able condensers. sembly is illustrated below.

AFTER more than two years of testing by the technical staff of Aluminum Company of Amer- ica, and by the designing engineers of the leading manufacturers of receiving sets, nine manufacturers have adopted Alcoa Radio Sheet for their condenser blades. Paralleling the increased use of Alcoa Radio Sheet are large in- In 1928 radio manufacturers creases in the use of aluminum used almost three times as much for shielding, aluminum foil for Alcoa Radio Sheet as was used in fixed condensers, and aluminum 1927, and more than six times as die castings for loud speaker much as in 1926. In1929 more than housings, chasses and condenser 6,000,000 single condenser units frames. will be made of Alcoa Radio Sheet. We will be glad to send you, on This wide and rapidly growing request, a copy of the booklet, use of Alcoa Radio Sheet is due to "Aluminum for Radio." its extreme accuracy of gauge, high electrical conductivity, unique Aluminum Company of America freedom from vibrating, its light- 2470 Oliver Building Pittsburgh. Pa. ness and its workability. Offices in 18 Principal American Cities Alcoa Radio Sheet, the exclusive product of Aluminum Company of America, is manufactured to limits of tolerance and uniformity hitherto unattainable.Its maximumtotalvariation withinasingle sheet is .0005 inch. Its sheet to sheet tolerance is +.00I inch. Itis patent leveled, highly planished, and accurately Aiired. We will be glad to quote on finished blades of high accuracy made from Alcoa Radio Sheet. ALUMINUM `Me mat of quality in Radio

When writing to advertisers mention of the PROCEEDINGS will be mutually helpful. XXVIII Standard of Accuracy and Long Life SJ ROYALTY Variable High Resistances Exclusively Licensed by Technidyne Corporation Under U.S. Patents No. 1593658-1034103-1034104

ROYALTY Variable High Resistances have estab- lishedthemselvesasaforemostchoicewith manufacturers and set builders because of their uncom- mon accuracy, efficiency, smoothness of operation, and dependability. They are widely usedas volume controls and are especially adapted for use in high frequency circuits where itis important that resistances which are free from harmful inductance and capacity effects be used. These resistances are made with the best insulating materials.The resistance element is under control in manufacture and does not change in use.There is no binding-knob and shaft turn smoothly over full range which is covered in one turn. There is a Royalty for every radio purpose. Our Engi- neering Departmentwill make recommendationsof units with special gradients for specific circuit purposes. Electrad specializes in a full line of Controls for all Radio Purposes, including Television. Write for Helpful Circuit and Resistance Data Dept. PE3 175 Varick Street, New York ELECTRAD !l 1ien writing to advertisers mention of thePROCEEDINGS stillhe mutually helpful. XXIX Continental Resistors

Durable dependable, simple in structure and givea minimum of resistor trouble. Type A for grid leaks and light power purposes.Will A dissipate 1/2 watt safely.

Types W and X for greater 4Q1111111iJpower dissipation. N.^/

All types furnished in any resistance value desired.

In use continuously for a number ofyears by the largest manufacturers.

Types E2 and D2 furnished with wire leads solderedtocoppered ends, are for soldering permanently into posi- D2 tion in apparatus where they are to be used.

Samples for test sent on receipt of specifications. CONTINENTAL CARBON INC. WEST PARK, CLEVELAND, OHIO

When writing to advertisers mention of the PROCEEDINGS will be mutually helpful. XXX On Commander Byrd's Antarctic Expedition -another tribute to the Only DURHAMS are Used' DURHAMMetallized principle!-another tribute to the extreme care with which DURHAM Resistors, Powerohms and Suppressors are made!-another tribute to DURHAM accuracy and utter dependability!-read the above letter from Chief Radio Engineer Malcolm P. Hanson of the Byrd Antartic Expedition.In effect he says "We are using DURHAMS exclusively because past experience has taught us that they can be relied upon for perfect performance under even the most adverse conditions." DURHAM Resistances are available for every practical resistance purpose in radio and television work from 250 ohms to 100 Megohms and in ratings for all limited power purposes.Used in leading radio laboratories, endorsed by leading engineers and sold by leading job- bers and dealers.Descriptive literature on the entire line of DUR- HAM products will be gladly sent upon request.

RESISTORS & POWEROHMS International Resistance Company, 2006 Chestnut Street, Philadelphia, Pa.

II. hen tcwiting to adlwitisc, s mention of thePRocEEniN6swill be mut tmll2.. hr!p I lit XXXI FARADON exemplifies excellence in workmanship and materials

WIRELESS SPECIALTY APPARATUS COMPANY Jamaica Plain, Boston (Eat. 1907) Manufacturers o/

ELECTROSTATIC CONDENSERS FOR ALL PURPOSES 2690

Whenwriting to adzertisers mention of the PROCEEDINGS Win be mutually helpful. XXXII The New TEMPLE DYNAMIC Speakers

ADD totheapproved and accepted principle ofsoundreproductionthe compelling significance of the Temple name and the result is a product which again sets a new standard in speaker excellence.

Temple Dynamics are made onlyas Temple can make them-that means better.

Available to manufacturers in "They Speak three chassis models: for Model 10, 110 volt, A.C., 60 cycle Model 12, 110 volt, A.C., 25 cycle Themselves" Model 14, 110 volt, D.C.

Write for full particulars

TEMPLE CORPORATION 1925 S. Western Ave., Chicago, U. S. A. iMM41010144LEADERS61SMISNOWtSPEAKER IN DESIGN

When writing to advertisers mention of the PROCEEDINGS will be mutually helpful. XXXIII "Caveat Emptor (Let the Buyer Beware)

An ancient phrase ina dead language, but possessing a significance as potent todayas in the days of Imperial Rome, when it is believed to have originated.

There is no need to hesitate, however, ifitis Condensers you contemplate buying. When "canned" musicwas first being sent over the air and received with crystals and phones, thename CARDWELL was already identified with qualityapparatus and CARDWELL fair dealing was a "buy" word. Fads and Fancies come andgo, fostered and nurtured in the mass mind by not disinterested Editorial publicity.Some persist, some die, depending upon how long the Public respondsto the stimulus of the "ultimate" in Radio, goaded by flowerypromises of remarkable results from this or that. Uninfluenced by extravagant claims, guided only bythe calm, deliberate, coldly scientific conclusions reached byResearch En- gineers, such organizations as General Electric,Westinghouse, RCA, and Western Electric use CARDWELL CONDENSERSof many descriptions.Upon what do you base your conclusions when you buy condensers? Follow the leaders and be safe.Buy as they buy and you won't be far wrong.

CARDWELL CONDENSERS TRANSMITTING-RECEIVING k's.0 Za" "The Standard of Comparison" Send for Literature THE ALLEN D. CARDWELL MFG. CORPN. 81 Prospect St. BROOKLYN, N. Y. 11 71; n erritinyto advertisers mention of the PROCEEDINGS willbe mutuallyhelpful. XXXIV Superior resources of research and manufacturing guarantee to RCA Radiotrons the finest possible quality in vacuum tubes. They are the standard of the industry -and so accepted by both the trade and the public.

RADIO CORPORATION OF AMERICA NEW YORK CHICAGO ATLANTA DALLAS SAN FRANCISCO IIRWA 114/11D11 CITRON M .111E UT THE MAKERS OF Till RA DIOLA

When writing to advertisers mention of thePROCEEDINGSwill be mutually helpful. XXXV Only Centralab Has Resistances Like These THE construction and design of a variable resistance is of as great importance as the mere fact that it possesses a certain resistance and will carry a specified current load.Centralab design is such that the resistance unit not only will handle the power but also vary it in a manner so as to derive the greatest efficiency from the receiving set or power unit. The following features distinguish Centralab variable resistances of the Graphite Disc type: Rocking Disc Contact. (varies the graphite resistance by means of a pressure shoe, eliminating sliding contacts to wear out and become noisy.) One turn of knob gives complete variation. (the entire resistance range from maximum to minimum is obtained by one turn of the knob.) Insulated shaft and bushing. (eliminates body capacity in a critical circuit.) Constant resistance. (itis completely encased and cannot be affected by atmosphere or temperature changes.) Noiseless, smooth and easy adjustment. (no scraping or scratching; knob turns like velvet;can never bind or work hard.Absolutely no noise in the speaker.) Rigidly built; fully guaranteed. (allbakelite and special metal construction makes possible afool proof unit and an unqualified guarantee of satisfaction.) Made in two and three terminal units to be used as Volume Controls, Modulators, and Potentiometers.Special resistance tapers can be had for any circuit.

CENTRAL RADIO LABORATORIES 16 Keefe Ave. Milwaukee, Wia. A CENTRALAB CONTROL IMPROVES THE SET

If'hen writing to advertiser's mention of the PROCEEDINGSwi!1 be mutually helpful. XXXVI ciheWINDING °PERKY -0)1)1

IN THE manufacture of millions of by-pass and filter condensers, the Fast organization has had to unwind many a knotty problem. In overcoming one of the chief difficulties of production, it was found that specially designed winding machines could so be utilized as to insure an absolutely uniform product, at the same time enabling us to speed up production requirements tremendously. When you consider that orders pour into us from manufacturers, practically all at one time, you can see that exceptional production facilities must be highly specialized in order to meet such requirements. Whether your needs call for thousand lots or millions, you can depend upon it that the Fast Organization has the facilities for filling them quickly and economically.

Send us your specifications for all your needs or as an auxiliary source of supply.

11411l41Ili Jaw R AA A 3982 Barry Avenue, Chicago, Illinois I When writing to advertisers mention of thePROCEEDINGSwill be rnutually helpful. XXXVII and PERFECT PERFORMANCE FOLLOWS T. C. A. power transformers, like T. C. A. audio transformers,outputtransformersfor dynamic speakers, and chokes, are built to standards critical enough to meet the most exacting in the industry. For among T. C. A. users are manufac- turersoftheworld'sfinest radio sets who must have the best. Refinements are many: automatically wound coils, vacuum impregnated; tinnedleads;smoothlaminations free from burrs; and perfect insulation. These and other important improvementsinconstructionnot onlyinsure good performance but save the manufacturer money on his assembly. Made up to suit individual specifications in separate units, complete assemblies, or semi -mounted. These quality products cost no more than you have been paying, and you gain the perfect performance which invariably follows their use. The Transformer Corporation of America 1428-1432 Orleans Street CHICAGO, ILL. Sales Offices in Principal Cities At,. ios Intermediate and pushpull stages

DEPENDABLE VOLUME DELIVERIES Output because of transformers COMPLETE MANUFACTURING Chokes RESOURCES

When writing to advertisers mention of the PROCEEDINGS will be mutually helpful. XXXVIII CHECK...CHECK...CHECK...CHECK through a barrage of inspections

Examined at every step that have spelled success for ... checked at every opera- ARCTURUSusers ... that tion ... tested at frequent have madeARCTURUS intervals...that is the lot Tubes the basis by which of eachARCTURUS A -C other tubes are judged. Tube... interminably The engineering attain- "on trial." ments inARCTURUS A -C Not a tube escapes.It Tubes are sound reasons must measure up to the why critical engineers and most rigid standards set by manufacturers demand our engineers.Standards these Long -Life blue tubes.

efENGINEERING FACTS HAVE A UTILITY SIGNIFICANCE TO THE BROADCAST LISTENER ARCTURUSARCTURUSRADIO CO. NEWARK, N.J. A LONG LIFETUBESII

When writing to advertisers mention of thePROCEEDINGSwill be mutually XXXIX Helping WRNI? to keep illsfinere utation

STATION WRNY, Hotel Roosevelt, New York, is known for the clarity of its recording and transmissions. The engineers here, as in other large broadcasting stations, depend upon PYREX Insulators as an essential to long range and protection of tone quality against retransmission noises from adjacent conductors. If you want the best transmission or improved reception and one thing less on the trouble list, equip your antenna and lines with PYREX In- sulators.The insulating qualities, the mechanical strength, the super- hard smooth time -and -element -resisting surface, and the resistance to destruction originate in the molten glass and are imperishable. Correct antenna, strain, entering, stand-off, pillar and bus bar types are easily chosen from our booklet, f-PYRRadiE2C "PYREX Radio Insulators." Insulators Write to us for a free copy of the booklet and get PYREX Insulators from your supply house. Yid CORNING GLASS WORKS, Dept. 63 Industrial and Laboratory Division CORNING, N. Y. PYREX RADIO INSULATORS ....1 Product of Corning Glass Itbrks

When writing to advertisersIIIntion of thePROCEEDINGS willbe mutually helpful. XL TO ALL DYNAMIC SPEAKER jtIANUFACTURERS: A New Rectifier Development

ELKON, INC. announces the development of a new type of rectifier for dynamic speakers which represents a notable improvement over the old types. This latest development of the Elkon laboratories has been released only after exhaustive tests. The Elkon Equipped Dynamic Speaker is humless and costs less to produce. Installations of this new Elkon rectifier-for your approval-will be made on all dynamic speakers sent to our laboratories. Samples furnished to representative manufacturers. Or, if you prefer, one of our engineers will call at the plant.

ELKON, Inc. Division P. R. Mallory & Co., Inc., 350 Madison Ave.,New York nAn441,44....era..a.ra.nai.ar/raaenathenathArAnAnbaathAreannnenIntAnaahnara. Whim writing to advertisers mention of thePROCEEDINGS willbe mutually helpful. X1.1 'It's great, Bill, since you removed the ADENOIDS'" -And indeed it is. Bill knows the differencenow between ordinary "radioed" music and that which the AmerTran Power Amplifier and Hi -Power Box gives him. And his friends do, as well. Bill performed the "adenoid" operation on his set-and when you do-you will know what radio can be like. Take out the inferior audio system-replace it with the best that money can buy-and your set, no matter how old or out of date will be better in tone than the most expensive receivers on the market. The new AmerTran Power Amplifier push-pull for 210 tubes and the improved ABC Hi -Power Box will do the trick or if you do not want to spend that much money use the push-pull amplifier for 171 tubes or a pair of AmerTran De Luxe transformers. Any AmerTran outfit will eliminate the adenoids. See your dealer or write to us today.

AmerTran ABC Hi -Power Box -500 volts DC plate volt- _Maw age, current up to 110 ma; AC filament current for all tubes for any set. Adjustable AMERICAN TRANSFORMER COMPANY bias voltages for alltubes. Transformer Manufacturers for more than 29 years Price, east of Rockies-less tutees-$95.W. 82 Emmet St. Newark, N. J.

When writing to advertisers mention of the PROCEEDINGS Will be mutuallyhelpful. XLII Ready Soon! The NEW UNITED SCIENTIFIC VARIABLE CONDENSERS

WRITE FOR GENERAL SPECIFICATIONS

1:141174P.,:4111174.,,,s7411.:40,wesr,-nsv...6.-terav-41....a.-11;:sso:sliw.svi.:411.,-ss.7.;:rs.7.4111;a1.7411Villr7.49.70

UNITED SCIENTIFIC LABORATORIES, INC. 115-C Fourth Avenue, New York City

Branch Offices for Your Convenience in ro St. Loins Boston Cincinnati Philadelphia London, Ontario Chicago Minneapolis Los Angeles San Francisco

41111;" 411. -41. 0 1144 "'411r

Whenwriting to advertisers mention of thePROCEEDINGS willbe mutually' helpful. XLIII Jewell Miniature Instruments In Moulded Bakelite Cases

NOW the JewellMiniature Movements,whichhave proved sosatisactorytoradio manufacturers for shop testing service, as well as servicing equipment, are available in moulded bakelite cases at no extra cost.In addition to high insulating value, No. 68 the bakelite cases provide a permanent high finish-there is no enamel to wear off.

A REFINED MOVEMENT Jewell Patterns No. 68, 78, and 88 are equipped with a narrower pointer,resultinginincreased damping and doubling the speed of action, an important factor for No. 78 shop service.Consequently these new instruments in bakelite cases provide quicker and more accurate reading.It will pay you to write for folder describing these instruments, or ask the Jewell representative in your territory.

Jewell Electrical Instrument Co. 1650Walnut St., Chicago, Illinois No. 88

11 GOOD INSTRUMENTS 11111 "DI I Radio Instruments11111111 11 hen writing to advertisers mention of the PROCEEDINGS will be mutually helpful. XLIV Choose Your

Here are showr fourteenpopular types of wire and insulations eachmade in E full range of commercial sizes. I. Silk Covered Enam- R.Bare or Tinned eied Wire Coster Wire 2. Silk Covered Magnet Squc.re and Rec- Wire tangularMagnet 3. "Ducotto Wire." A Wire cotton covered wire, Stranded anc enameled Braided Cables EnameledMagnet 4. Cotton Covered En- Wire ameledWire Flexible Coll Leads 5. Cotton Covered Mag- Antenna Wire- net Wire Stranded 6. H Igh Frequenc7 Ca- Enameled or bare bles. l.itiendrala wire Squareand Rec- 7. Flat Braided Cables tangular Wire Special wires madetospecifications-the same high Dudlo standards always rigidly maintained. 8 DUDLO 0

DUDLO MANUFACTURING CO., FORT WAYNE, INDIANA Division of General Cable Corporation 420 Lexington Ave. 105 West Adams St. 274 BrannanSt. 4143 Bingham Ave. NEW YORK CITY CHICAGO, ILL. SAN FRANCISCO, CALIF. ST. LOUIS, MO.

When writing to advertisers mention of thePROCEEDINGSwill be mutually helpful. X LV PROFESSIONAL ENGINEERING DIRECTORY For Consultants in Radio and Allied Engineering Fields

The J. G. White Electrical Testing Engineering Corporation Laboratories RADIO DEPARTMENT Engineers-Constructors also Builders of New York Radio Electrical, Photometric, Central Chemical and Mechanical Industrial, Steam Power, and Gas PlantsSteamandElectricRail - Laboratories rods, Transmission Systems. 80th Street and East End Ave. 43 Exchange Place New York NEW YORK, N. Y. Amy, Aceves & King, Inc. PATENTS Consulting Engineers WM. G. H. FINCH DESIGN-TEST-DEVELOPMENT Radio Transmitters, Receivers and Sound Patent Attorney Reproducing Apparatus ( Registered U. S. & Canada) Research Laboratories Mem. I. R. E. Mem. A. I. E. E. 55 West 42nd Street, New York 303 Fifth Ave. New York Longacre 8579 Caledonia 5331 For EXPORT of material made by: ROBERT S. KRUSE Allen D. Cardwell Mfg. Corp. Claroetat Mfg. Co. Consultant for development of Corning Glass Works Short-wave Devices Dubilier Condenser Corp. Elkon, Inc. Radio Engineering Labs. 103 Meadowbrook Road Raytheon Mfg. Co. United Scientif.c Labs. WEST HARTFORD, CONN. Weston Electrical Inst. Corp. apply to Telephone, Hartford 45327 AD. AURIEMA, INC. 116 Broad Street, New York, N.Y.

JOHN MINTON, Ph.D. Consulting Engineer for BRUNSON S. McCUTCHEN Developing - Designing - Manufacturing Consulting Radio Engineer of Radio Receivers, Amplifiers, Transform- 17 State Street ers,Rectifiers,SoundRecordingand Reproducing Apparatus. NEW YORK Radio and Electro-Acoustical Laboratory 8 Church St. White Plains, N. Y.

When writingto advertisers mention ofhePR, .1.- ELM SOS winbe mutually helpful. XLVI PROFESSIONAL ENGINEERING DIRECTORY For Consultants in Radio and Allied Engineering Fields (Continued)

4th Edition Complete Line of Thoroughly Revised-Up-to-Date RADIO PANELS, TUBING, "RADIO THEORY AND RODS AND INSULATING OPERATING" MATERIALS Drilling,Machining andEngravingto 992 Pages800 Illustrations Specifications By MARY TEXANNA LOOMIS TELEVISION KITS, DISCS, NEON President and Lecturer on Radio TUBES, PHOTO ELECTRIC CELLS Loomis Radio College AND ALL PARTS Member Institute of Radio Engineers Write for Catalog Price$3.50-Postage Paid Insuline Corp. of America LOOMIS PUBLISHING CO. 78-80 Cortlandt St., N. Y. Dept. 6, Washington, D.C. Cortlandt 0880

DODGE RADIO SHORTKUT Ramsey's Experimental Radio With Appendix and Hints for Better Key Work.Fixes Signals in mind to Third Edition stick-Kills Hesitation, Cultivates Code (xii-229 pages,51/2x71/4,clot}, Reading,Speed, and Good Fist-Pro- 152 figures) duces Results. Slow Hams raise speed to 25 per in few evenings. Previous Failures A Radio Frequency Manual qualify and pass exam quickly.Begin- Fundamental measurements and ners master code and pass in ten days. tests. REPORTS FROM USERS Review in June 28 Proc. Inst. Rad. Eng. Tellthe complete story-Mailed on p. 851 request. Price$3.50. Money order. Price $2.75 postpaid None C. 0. D.Foreign add 50 cents. Ramsey Publishing Co. C. DODGE, MAMARONECK, N.. Y. 615 E. 3rd St., Bloomington, Indiana

ENGINEERS WANTED Large independent Eastern manufacturer EUROPE of thermionic vacuum tubes requires serv- AMERICAN RADIO MAN ices of several engineers.Experience in tube manufacture desirable but not abso- of wide experienceintesting,buying, lutely necessary.Applicants should be selling,manufacture, andpatents,with collegegraduates. Give education,ex- connections throughout Europe and own plantinBerlin undertakes commissions perience, salary required, and other per- in all matters pertaining to radio art and tinent information infirst letter. trade. Address replies to Box 826, I.R.E. "Kondax," Hutten Str. 31, Berlin NW. 87

RADIO ENGINEERS Your card on this new professional card page will give you a direct introduction to over 6,000 technical men, executives, and others with important radio interests. Per Year (12 issues) $120.00

When writingto u.k.ertisers mention of the PROCEEDINGS will be mutually helpful. XLVII Alphabetical Index to Advertisements A Aerovox Wireless Corp XV Allen-Bradley Co XVIII Aluminum Co. of America XXVIII American Transformer Co. XLII Arcturus Radio Ce XXXIX B Braun, W. C., Co XXVII C Cardwell, Allen D. Mfg. Corp. XXXIV CeCo Mfg. Co., Inc XXVI Central Radio Laboratories XXXVI Clarostat Mfg. Co., Inc XXV Cohn, Sigmund V Condenser Corp. of America XXIII Continental Carbon Inc. XXX Corning Glass Works XL D De Jur-Amsco Corp ...LI Dubilier Condenser Corp L Dudlo Manufacturing Company XLV E Eby, H. H., Mfg. Co., Inc., The XXII Electrad, Inc XXIX Electric Specialty Co VIII Elkon, Inc XLI F Fast, John E. & Co XXXVII Formica Insulation Co. XXIV G General Radio Co Outside Back Cover Grebe, A. H. and Co., Inc Inside Back Cover H Hammarlund Mfg. Co I XXI International Resistance Co X X XI I.R. E XVI Isolantite Co. of America VI, VII J Jewell Electrical Instrument Co XLIV K Karas Electric Company XVII L Lentz, C. R., Inc XIII M Mertz Specialty Co LII N National Vulcanized Fibre Co X P Pacent Electric Co., Inc n Polymet Manufacturing Corp XI Potter Co., The XII Professional Engineering Directory , XLVI, XLVII R Radio Corporation of America XXXV Radio Engineering Laboratories XLIX Radio Receptor Co. IX Raytheon Mfg. Co IV Roller -Smith Co XX S Scovill Manufacturing Co XIV Silver -Marshall, Inc III T Temple Corporation XXXIII Thordarson Electric Mfg. Co I Transformer Corporation of America, The XXXVIII U United Scientific Laboratories, Inc. XLIII V Valley Appliances, Inc XIX W Wireless Specialty Apparatus Co XXXII

When writing to advertisersmentionof the PROCEEDINGS will be mutually helpful. XLVIII PRECISION WAVEMETERS

CAT. NO. 142

An accurately calibrated wavemeter, reasonably priced, and suitable for all General Laboratory Purposes. Rugged straightline wavelength condenser- Vernierindicating dial-dialscaleaccurately divided in 1000 parts-assures micrometer readings-coils wound on heavy grooved Bakelite forms-windings cannot be damaged-indepen- dentresonanceindicatingcircuit-comprises special pick up coil, D.C. milliammeter and rec- tifying crystal-large calibrationcurves with duplicatecopies-calibratedagainststandard Piezo oscillator to guaranteed 0.2% accuracy- housed in walnut case having removable cover. SUPPLIED IN TWO SIZES Type "B"- 10 to 100 Meters Price $65.00 Type "C"-100 to 600 Meters Price $65.00 .sr, ,or MANUFACTURES A COMPLETE LINE OF APPARATUS FOR SHORT WAVE TRANS- MISSION AND RECEPTION. s -Or -lor RADIO ENGINEERING LABORATORIES 100 Wilbur Ave. Long Island City, N.Y., U.S.A.

ILIwn writing to aivertisers mention of thePROCEEDINGSwill be mutually helpful. XL1X You Can Forget the Condensers, If They Are DUBILIER'S

TYPE 596 B For high frequency furnace- tube bombarders-radio franinittere.etc.

There is no substitute for Quality Since 1913 Dubilier has been producing condensers of many and varied types filling every need in the radio field. Mica condensers...Paper condensers ...Transmitting condensers... are but afew of the many hundred types. Ever since the event ofRadio,Dubilier has been the manufacturers' standard -and the set builders' stand-by. Built in every Dubilier Condenser is a factor of safety which is your safe- guard for years of service without failure. Write Dept. 62 for Free Catalog

Type PL -1152 especially designed for the Thordarson 250-2 Stage Power Ampli- Dubilier fier and Plate supply,CONDENSER CORPORATION and Thordarson 250 Plug in Power and 10 East 43rd Street, New York City Plate Supply. Used with Thordarson T- Address for free 2900 Power Supply Dept. 84 catalog Transformer. Price $17.50. Res.U.SPet_Off.

When writing to advertisers mention of thePROCEEDINGS Willhe mutually helpful. DOEINSCO PARTS SPECIALISTS FOR SET MANUFACTURERS

QUANTITY

Neutralizing Condensers-All PRODUCTION Capacities OF QUALTITY Tip Jacks-Plain and Insulated PRODUCTS DELIVERED Resistances That Do Not Change ON

=10-1_01 SCHEDULE

Condensers in All Capacities and Gangs-Accurately Let Us Quote on Your Matched Specifications DeuR-Amsco coRp Resistance and Condenser Headquarters Broome & Lafayette Sts., New York City

When writing to advertisers mention of the PROCEEDINGS will be mutually helpful. LI Mertz Nophonic Tube Caps

FIRST SECOND For large set manufac- Fortheretailtrade turers,canbefur- and small set manufac- nishedplainor em- turer is sold embossed bossed with your name MertzSpec.Co.at ortrademark,in 25¢ list.Types 200 weights from two and 199; weight 5 and ounces to eight ounces. 21Z ounces respectively.

Saves servicing, complaints, and loss of good will for the set manufacturer, jobber and dealer.Write for prices and dis- counts.

MERTZ "SADDLE"

GROUND CLAMPS

The Clamp That Stays Put Anywhere Strap Phosphor Bronze, Rest Brass Made in 1", 2", and 3" Sizes

We are developing for set manufacturers individual aluminum shields; for ordinary tubes, grid shield tubes, and transformers. Write for details. MERTZ SPECIALTY COMPANY 1306 Stockton Street WILMINGTON, DELAWARE

When writing to advertisers mention of the PROCEEDINGS will be mutually helpful. LII View of Grebe CR-18 Special Short WaveReceiver Short Wave Exploration VOR RADIO enthusiasts who liketo drop 2- in on London, the Coast and the Byrd expedition two or three eveningsa week, we suggest a CR-18 Special. In the 20-40 . i / meter band, in which the Byrd expedition is usually picked up, the new Special has \\ greater radio frequency efficiency. ----, \ Provides two stages of audio, making the _----.'-- (/----- CR-18 Special adaptable to theuse of a power tube. Also permits loud speaker operation. Volume control-from headset ---4t level to full audio output-bya single 't adjustment. Write us for more details. i---:," i A. 11. GREBE & CO., INC. Richmond Hill, N. Y. ,),.,....,. ji' Western Branch: 443 So. San Pedro St., Los Angeles, Calif.

GEC%mowSYNC, IRorltAsj., RADIO MAKERS OF' QUALITYRADIO SINCE 190

11' hen writing to advertisers mention of the PROCEEDINGS will be NIU/14(1//y help! id. Station Frequency Meter

Type 532

The Type 532 Station Frequency Meter offers a frequency standard of the convenient resonant -circuit type which may be set and read to within 20 cycles. By means of a condenser of high minimum capacity and small variation, the entire scale of the instrument is used for a very narrow band of frequencies, centered on the station frequency. This feature in itself permits a very close setting of the instrument, but it is supplemented by a novel device which permits setting to the peak of resonance curve with remarkable precision.The absolute accuracy of the calibration cannot be guaranteed to be within the precision of setting.The Type 532 Meters are supplied with a six -months' guarantee of accuracy of the scale division corresponding to the station frequency within 500 cycles. Further Details on Request GENERAL RADIO COMPANY 30 STATE STREET CAMBRIDGE, MASS.

OSOSCIE SANTA PUFLISNING COMPANY MENASHA, WISCONSIN